Online planning and execution of minimum-time maneuvers on three-dimensional (3D) circuits is an open challenge in autonomous vehicle racing. In this paper, we present an artificial race driver (ARD) to learn the vehicle dynamics, plan and execute minimum-time maneuvers on a 3D track. ARD integrates a novel kineto-dynamical (KD) vehicle model for trajectory planning with economic nonlinear model predictive control (E-NMPC). We use a high-fidelity vehicle simulator (VS) to compare the closed-loop ARD results with a minimum-lap-time optimal control problem (MLT-VS), solved offline with the same VS. Our ARD sets lap times close to the MLT-VS, and the new KD model outperforms a literature benchmark. Finally, we study the vehicle trajectories, to assess the re-planning capabilities of ARD under execution errors. A video with the main results is available as supplementary material.

Most autonomous cars rely on the availability of high-definition (HD) maps. Current research aims to address this constraint by directly predicting HD map elements from onboard sensors and reasoning about the relationships between the predicted map and traffic elements. Despite recent advancements, the coherent online construction of HD maps remains a challenging endeavor, as it necessitates modeling the high complexity of road topologies in a unified and consistent manner. To address this challenge, we propose a coherent approach to predict lane segments and their corresponding topology, as well as road boundaries, all by leveraging prior map information represented by commonly available standard-definition (SD) maps. We propose a network architecture, which leverages hybrid lane segment encodings comprising prior information and denoising techniques to enhance training stability and performance. Furthermore, we facilitate past frames for temporal consistency. Our experimental evaluation demonstrates that our approach outperforms previous methods by a large margin, highlighting the benefits of our modeling scheme.

We present a novel autonomous driving framework, DualAD, designed to imitate human reasoning during driving. DualAD comprises two layers: a rule-based motion planner at the bottom layer that handles routine driving tasks requiring minimal reasoning, and an upper layer featuring a rule-based text encoder that converts driving scenarios from absolute states into text description. This text is then processed by a large language model (LLM) to make driving decisions. The upper layer intervenes in the bottom layer’s decisions when potential danger is detected, mimicking human reasoning in critical situations. Closed-loop experiments demonstrate that DualAD, using a zero-shot pre-trained model, significantly outperforms both rule-based and learning-based motion planners when interacting with reactive agents. Our experiments also highlight the effectiveness of the text encoder, which considerably enhances the model’s scenario understanding. Additionally, the integrated DualAD model improves with stronger LLMs, indicating the framework’s potential for further enhancement. Code and benchmarks are available at github.com/TUM-AVS/DualAD.

Autonomous drifting is a complex and crucial maneuver for safety-critical scenarios like slippery roads and emergency collision avoidance, requiring precise motion planning and control. Traditional motion planning methods often struggle with the high instability and unpredictability of drifting, particularly when operating at high speeds. Recent learning-based approaches have attempted to tackle this issue but often rely on expert knowledge or have limited exploration capabilities. Additionally, they do not effectively address safety concerns during learning and deployment. To overcome these limitations, we propose a novel Safe Reinforcement Learning (RL)-based motion planner for autonomous drifting. Our approach integrates an RL agent with model-based drift dynamics to determine desired drift motion states, while incorporating a Predictive Safety Filter (PSF) that adjusts the agent’s actions online to prevent unsafe states. This ensures safe and efficient learning, and stable drift operation. We validate the effectiveness of our method through simulations on a Matlab-Carsim platform, demonstrating significant improvements in drift performance, reduced tracking errors, and computational efficiency compared to traditional methods. This strategy promises to extend the capabilities of autonomous vehicles in safety-critical maneuvers.

This work presents a methodology to estimate tire parameters and their uncertainty using a Bayesian optimization approach. The literature mainly considers the estimation of tire parameters but lacks an evaluation of the parameter identification quality and the required slip ratios for an adequate model fit. Therefore, we examine the use of Stochastical Variational Inference as a methodology to estimate both - the parameters and their uncertainties. We evaluate the method compared to a state-of-the-art Neider-Mead algorithm for theoretical and real-world application. The theoretical study considers parameter fitting at different slip ratios to evaluate the required excitation for an adequate fitting of each parameter. The results are compared to a sensitivity analysis for a Pacejka Magic Formula tire model. We show the application of the algorithm on real-world data acquired during the Abu Dhabi Autonomous Racing League and highlight the uncertainties in identifying the curvature and shape parameters due to insufficient excitation. The gathered insights can help assess the acquired data's limitations and instead utilize standardized parameters until higher slip ratios are captured. We show that our proposed method can be used to assess the mean values and the uncertainties of tire model parameters in real-world conditions and derive actions for the tire modeling based on our simulative study.

Developing robust, efficient navigation algorithms is challenging. Rule-based methods offer interpretability and modularity but struggle with learning from large datasets, while end-to-end neural networks excel in learning but lack transparency and modularity. In this paper, we present MIND-Stack, a modular software stack consisting of a localization network and a Stanley Controller with intermediate human interpretable state representations and end-to-end differentiability. Our approach enables the upstream localization module to reduce the downstream control error, extending its role beyond state estimation. Unlike existing research on differentiable algorithms that either lack modules of the autonomous stack to span from sensor input to actuator output or real-world implementation, MIND-Stack offers both capabilities. We conduct experiments that demonstrate the ability of the localization module to reduce the downstream control loss through its end-to-end differentiability while offering better performance than state-of-the-art algorithms based on traditional path tracking approaches. We showcase sim-to-real capabilities by deploying the algorithm on a real-world embedded autonomous platform with limited computation power and demonstrate simultaneous training of both the localization and controller towards one goal. While MIND-Stack shows good results, we discuss the incorporation of additional modules from the autonomous navigation pipeline in the future, promising even greater stability and performance in the next iterations of the framework.

This work presents an online velocity planner for autonomous racing that adapts to changing dynamic constraints, such as grip variations from tire temperature changes and rubber accumulation. The method combines a forward-backward solver for online velocity optimization with a novel spatial sampling strategy for local trajectory planning, utilizing a three-dimensional track representation. The computed velocity profile serves as a reference for the local planner, ensuring adaptability to environmental and vehicle dynamics. We demonstrate the approach's robust performance and computational efficiency in racing scenarios and discuss its limitations, including sensitivity to deviations from the predefined racing line and high jerk characteristics of the velocity profile.

Autonomous vehicles (AVs) must navigate dynamic urban environments where occlusions and perception limitations introduce significant uncertainties. This research builds upon and extends existing approaches in risk-aware motion planning and occlusion tracking to address these challenges. While prior studies have developed individual methods for occlusion tracking and risk assessment, a comprehensive method integrating these techniques has not been fully explored. We, therefore, enhance a phantom agent-centric model by incorporating sequential reasoning to track occluded areas and predict potential hazards. Our model enables realistic scenario representation and context-aware risk evaluation by modeling diverse phantom agents, each with distinct behavior profiles. Simulations demonstrate that the proposed approach improves situational awareness and balances proactive safety with efficient traffic flow. While these results underline the potential of our method, validation in real-world scenarios is necessary to confirm its feasibility and generalizability. By utilizing and advancing established methodologies, this work contributes to safer and more reliable AV planning in complex urban environments. To support further research, our method is available as open-source software at https://github.com/TUM-AVS/OcclusionAwareMotionPlanning.

Conventional trajectory planning approaches for autonomous vehicles often assume a fixed vehicle model that remains constant regardless of the vehicle's location. This overlooks the critical fact that the tires and the surface are the two force-transmitting partners in vehicle dynamics; while the tires stay with the vehicle, surface conditions vary with location. Recognizing these challenges, this paper presents a novel framework for spatially resolving dynamic constraints in both offline and online planning algorithms applied to autonomous racing. We introduce the GripMap concept, which provides a spatial resolution of vehicle dynamic constraints in the Frenet frame, allowing adaptation to locally varying grip conditions. This enables compensation for location-specific effects, more efficient vehicle behavior, and increased safety, unattainable with spatially invariant vehicle models. The focus is on low storage demand and quick access through perfect hashing. This framework proved advantageous in real-world applications in the presented form. Experiments inspired by autonomous racing demonstrate its effectiveness. In future work, this framework can serve as a foundational layer for developing future interpretable learning algorithms that adjust to varying grip conditions in real-time.

The classical g-g diagram, representing the achievable acceleration space for a vehicle, is commonly used as a constraint in trajectory planning and control due to its computational simplicity. To address non-planar road geometries, this concept can be extended to incorporate g-g constraints as a function of vehicle speed and vertical acceleration, commonly referred to as g-g-g-v diagrams. However, the estimation of g-g-g-v diagrams is an open problem. Existing simulation-based approaches struggle to isolate non-transient, open-loop stable states across all combinations of speed and acceleration, while optimization-based methods often require simplified vehicle equations and have potential convergence issues. In this paper, we present a novel, open-source, quasi-steady-state black box simulation approach that applies a virtual inertial force in the longitudinal direction. The method emulates the load conditions associated with a specified longitudinal acceleration while maintaining constant vehicle speed, enabling open-loop steering ramps in a purely QSS manner. Appropriate regulation of the ramp steer rate inherently mitigates transient vehicle dynamics when determining the maximum feasible lateral acceleration. Moreover, treating the vehicle model as a black box eliminates model mismatch issues, allowing the use of high-fidelity or proprietary vehicle dynamics models typically unsuited for optimization approaches. An open-source version of the proposed method is available at: https://github.com/TUM-AVS/GGGVDiagrams

Emergent-scene safety is the key milestone for fully autonomous driving, and reliable on-time prediction is essential to maintain safety in emergency scenarios. However, these emergency scenarios are long-tailed and hard to collect, which restricts the system from getting reliable predictions. In this paper, we build a new dataset, which aims at the longterm prediction with the inconspicuous state variation in history for the emergency event, named the Extro-Spective Prediction (ESP) problem. Based on the proposed dataset, a flexible feature encoder for ESP is introduced to various prediction methods as a seamless plug-in, and its consistent performance improvement underscores its efficacy. Furthermore, a new metric named clamped temporal error (CTE) is proposed to give a more comprehensive evaluation of prediction performance, especially in time-sensitive emergency events of subseconds. Interestingly, as our ESP features can be described in human-readable language naturally, the application of integrating into ChatGPT also shows huge potential. The ESP-dataset and all benchmarks are released at https://dingrui-wang.github.io/ESP-Dataset/.

This work aims to present a three-dimensional vehicle dynamics state estimation under varying signal quality. Few researchers have investigated the impact of three-dimensional road geometries on the state estimation and, thus, neglect road inclination and banking. Especially considering high velocities and accelerations, the literature does not address these effects. Therefore, we compare two- and three-dimensional state estimation schemes to outline the impact of road geometries. We use an Extended Kalman Filter with a point-mass motion model and extend it by an additional formulation of reference angles. Furthermore, virtual velocity measurements significantly improve the estimation of road angles and the vehicle’s side slip angle. We highlight the importance of steady estimations for vehicle motion control algorithms and demonstrate the challenges of degraded signal quality and Global Navigation Satellite System dropouts. The proposed adaptive covariance facilitates a smooth estimation and enables stable controller behavior. The developed state estimation has been deployed on a high-speed autonomous race car at various racetracks. Our findings indicate that our approach outperforms state-of-the-art vehicle dynamics state estimators and an industry-grade Inertial Navigation System. Further studies are needed to investigate the performance under varying track conditions and on other vehicle types.

Propagating uncertainties through nonlinear system dynamics in the context of Stochastic Nonlinear Model Predictive Control (SNMPC) is challenging, especially for high-dimensional systems requiring real-time control and operating under time-variant uncertainties such as autonomous vehicles. In this work, we propose an Adaptive SNMPC (aSNMPC) driven by Deep Reinforcement Learning (DRL) to optimize uncertainty handling, constraints robustification, feasibility, and closed-loop performance. To this end, our SNMPC uses Polynomial Chaos Expansion (PCE) for efficient uncertainty propagation, limits its propagation time through an Uncertainty Propagation Horizon (UPH), and transforms nonlinear chance constraints into robustified deterministic ones. We conceive a DRL agent to proactively anticipate upcoming control tasks and to dynamically reduce conservatism by determining the most suitable constraints robustification factor κ, and to enhance feasibility by choosing optimal UPH length Tu. We analyze the trained DRL agent’s decision-making process and highlight its ability to learn context-dependent optimal parameters. We showcase the enhanced robustness and feasibility of our DRL-driven aSNMPC through the real-time motion control task of an autonomous passenger vehicle when confronted with significant time-variant disturbances while achieving a minimum solution frequency of 110Hz. The code used in this research is publicly accessible as open-source software: https://github.com/bzarr/TUM-CONTROL

Current validation methods often rely on recorded data and basic functional checks, which may not be sufficient to encompass the scenarios an autonomous vehicle might encounter. In addition, there is a growing need for complex scenarios with changing vehicle interactions for comprehensive validation. This work introduces a novel synchronous multi-agent simulation framework for autonomous vehicles in interactive scenarios. Our approach creates an interactive scenario and incorporates publicly available edge-case scenarios wherein simulated vehicles are replaced by agents navigating to predefined destinations. We provide a platform that enables the integration of different autonomous driving planning methodologies and includes a set of evaluation metrics to assess autonomous driving behavior. Our study explores different planning setups and adjusts simulation complexity to test the framework’s adaptability and performance. Results highlight the critical role of simulating vehicle interactions to enhance autonomous driving systems. Our setup offers unique insights for developing advanced algorithms for complex driving tasks to accelerate future investigations and developments in this field. The multi-agent simulation framework is available as open-source software: https://github.com/TUM-AVS/Frenetix-Motion-Planner

Today’s software stacks for autonomous vehicles rely on HD maps to enable sufficient localization, accurate path planning, and reliable motion prediction. Recent developments have resulted in pipelines for the automated generation of HD maps to reduce manual efforts for creating and updating these HD maps. We present FlexMap Fusion, a methodology to automatically update and enhance existing HD vector maps using OpenStreetMap. Our approach is designed to enable the use of HD maps created from LiDAR and camera data within Autoware. The pipeline provides different functionalities: It provides the possibility to georeference both the point cloud map and the vector map using an RTK-corrected GNSS signal. Moreover, missing semantic attributes can be conflated from OpenStreetMap into the vector map. Differences between the HD map and OpenStreetMap are visualized for manual refinement by the user. In general, our findings indicate that our approach leads to reduced human labor during HD map generation, increases the scalability of the mapping pipeline, and improves the completeness and usability of the maps. The methodological choices may have resulted in limitations that arise especially at complex street structures, e.g., traffic islands. Therefore, more research is necessary to create efficient preprocessing algorithms and advancements in the dynamic adjustment of matching parameters. In order to build upon our work, our source code is available at https://github.com/TUMFTM/FlexMap_Fusion.

Our work introduces a module for assessing the trajectory safety of autonomous vehicles in dynamic environments marked by high uncertainty. We focus on occluded areas and occluded traffic participants with limited information about surrounding obstacles. To address this problem, we propose a software module that handles blind spots (BS) created by static and dynamic obstacles in urban environments. We identify potential occluded traffic participants, predict their movement, and assess the ego vehicle’s trajectory using various criticality metrics. The method offers a straightforward and modular integration into motion planning algorithms. We present critical real-world scenarios to evaluate our module and apply our approach to a publicly available trajectory planning algorithm. Our results demonstrate that safe yet efficient driving with occluded road users can be achieved by incorporating safety assessments into the planning process. The code used in this research is publicly available as open-source software and can be accessed at the following link: https://github.com/TUM-AVS/Frenetix-Occlusion.

This study introduces a novel approach to autonomous motion planning, informing an analytical algorithm with a reinforcement learning (RL) agent within a Frenet coordinate system. The combination directly addresses the challenges of adaptability and safety in autonomous driving. Motion planning algorithms are essential for navigating dynamic and complex scenarios. Traditional methods, however, lack the flexibility required for unpredictable environments, whereas machine learning techniques, particularly reinforcement learning (RL), offer adaptability but suffer from instability and a lack of explainability. Our unique solution synergizes the predictability and stability of traditional motion planning algorithms with the dynamic adaptability of RL, resulting in a system that efficiently manages complex situations and adapts to changing environmental conditions. Evaluation of our integrated approach shows a significant reduction in collisions, improved risk management, and improved goal success rates across multiple scenarios. The code used in this research is publicly available as open-source software and can be accessed at the following link: https://github.com/TUM-AVS/Frenetix-RL.

This research aims to investigate professional racing drivers’ expertise to develop an understanding of their cognitive and adaptive skills to create new autonomy algorithms. An expert interview study was conducted with 11 professional race drivers, data analysts, and racing instructors from across prominent racing leagues. The interviews were conducted using an exploratory, non-standardized expert interview format guided by a set of prepared questions. The study investigates drivers’ exploration strategies to reach their vehicle limits and contrasts them with the capabilities of state-of-the-art autonomous racing software stacks. Participants were questioned about the techniques and skills they have developed to quickly approach and maneuver at the vehicle limit, ultimately minimizing lap times. The analysis of the interviews was grounded in Mayring’s qualitative content analysis framework, which facilitated the organization of the data into multiple categories and subcategories. Our findings create insights into human behavior regarding reaching a vehicle’s limit and minimizing lap times. We conclude from the findings the development of new autonomy software modules that allow for more adaptive vehicle behavior. By emphasizing the distinct nuances between manual and autonomous driving techniques, the paper encourages further investigation into human drivers’ strategies to maximize their vehicles’ capabilities.

Determining the optimal cost function parameters of Model Predictive Control (MPC) to optimize multiple control objectives is a challenging and time-consuming task. Multi-objective Bayesian Optimization (BO) techniques solve this problem by determining a Pareto optimal parameter set for an MPC with static weights. However, a single parameter set may not deliver the most optimal closed-loop control performance when the context of the MPC operating conditions changes during its operation, urging the need to adapt the cost function weights at runtime. Deep Reinforcement Learning (RL) algorithms can automatically learn context-dependent optimal parameter sets and dynamically adapt for a Weights-varying MPC (WMPC). However, learning cost function weights from scratch in a continuous action space may lead to unsafe operating states. To solve this, we propose a novel approach limiting the RL action space within a safe learning space that we represent by a catalog of pre-optimized feasible BO Pareto-optimal weight sets. We conceive an RL agent not to learn in a continuous space but to select the most optimal discrete actions, each corresponding to a single set of Pareto optimal weights, by proactively anticipating upcoming control tasks in a context-dependent manner. This approach introduces a two-step optimization: (1) safety-critical with BO and (2) performance-driven with RL. Hence, even an untrained RL agent guarantees a safe and optimal performance. Simulation results demonstrate that an untrained RL-WMPC shows Pareto-optimal closed-loop behavior and training the RL-WMPC helps exhibit a performance beyond the Pareto-front. The code used in this research is publicly accessible as open-source software: https://github.com/bzarr/TUM-CONTROL

Modern robot systems, like autonomous vehicles, are complex, distributed systems that consist of many interacting components. End-to-end timing latency guarantees are key properties of such systems. They upper bound the data processing time and provide a predictable timing behavior. The Robot Operating System 2 (ROS 2) is a widely used and highly configurable set of software libraries for creating and deploying robot systems. It features a custom scheduler to execute time-triggered and event-triggered tasks and uses Data Distribution Services (DDS) for the communication between different system components. The data propagations between ROS 2 system components form cause-effect chains, which can be analyzed to determine the maximum reaction time (longest time between occurrence of an external cause and the earliest time when this external cause is fully processed) and maximum data age (longest time between the moment of a sensor measurement and the latest moment where an effect is based on this sensor measurement). In this paper, we provide an analysis of the end-to-end latencies in multi-executor ROS 2 systems to upper bound the end-to-end latencies of cause-effect chains in ROS 2 systems. Furthermore, we introduce an optimization using constrained programming that determines the optimal system configuration to minimize the end-to-end latencies for ROS 2 systems. We evaluate our upper-bound analysis to determine the end-to-end latencies of cause-effect chains in an autonomous driving-software stack for oval racing used in the Indy Autonomous Challenge and apply our optimization method to reduce the end-to-end latency upper bound, measured maximum, and measured mean by up to 50.2 %, 19.8 %, and 7.2 %, respectively.

The automotive industry is transitioning from traditional ECU-based systems to software-defined vehicles. A central role of this revolution is played by containers, lightweight virtualization technologies that enable the flexible consolidation of complex software applications on a common hardware platform. Despite their widespread adoption, the impact of containerization on fundamental real-time metrics such as end-to-end latency, communication jitter, as well as memory and CPU utilization has remained virtually unexplored. This paper presents a microservice architecture for a real-world autonomous driving application where containers isolate each service. Our comprehensive evaluation shows the benefits in terms of end-to-end latency of such a solution even over standard bare-Linux deployments. Specifically, in the case of the presented microservice architecture, the mean end-to-end latency can be improved by 5–8%. Also, the maximum latencies were significantly reduced using container deployment.

Robot localization is an inverse problem of finding a robot's pose using a map and sensor measurements. In recent years, Invertible Neural Networks (INN s) have successfully solved ambiguous inverse problems in various fields. This paper proposes a framework that approaches the localization problem with INN. We design a network that provides implicit map representation in the forward path and localization in the inverse path. By sampling the latent space in evaluation, Local_INN outputs robot poses with covariance, which can be used to estimate the uncertainty. We show that the localization performance of Local_INN is on par with current methods with much lower latency. We show detailed 2D and 3D map reconstruction from Local_INN using poses exterior to the training set. We also provide a global localization algorithm using Local_INN to tackle the kidnapping problem.

Real-time capability and robust software behavior have emerged as crucial issues since autonomous vehicles must react reliably to various traffic conditions when operating on our streets. The objective of our work is to understand and examine the processing latency of a software stack for autonomous vehicles. In this paper, we propose a framework based on ros2_tracing that automatically extracts implicit and explicit data flow from large-scale ROS 2-based autonomous driving software. It can measure the end-to-end latency and the individual components it is composed of. Using a static analysis, the implicit dependencies can be extracted. The method was used to analyze a software stack for autonomous vehicles. Compared to previous work that requires a manual definition of node-internal data dependencies and often does not follow the data flows completely, this paper provides a more feasible and comprehensive toolkit for analyzing real-world ROS 2 systems.

Violations of real-time properties and high latencies have emerged as crucial issues in autonomous vehicles since they can lead to unwanted vehicle behavior and critical maneuvers. Our study aims to provide a comprehensive understanding of latencies in a software stack for autonomous vehicles. In this paper, we present an evaluation workflow to inspect software and the occurring latencies for ROS 2 applications. This workflow was used to analyze the open-source autonomous driving stack Autoware. Universe by showing the influence of different soft- and hardware configurations. Our focus is on the evaluation of end-to-end, communication, computation, and idle latencies. Based on the results, we show the bottlenecks and motivate future directions to optimize ROS 2 autonomous driving software.

Autonomous vehicles demand high accuracy and robustness of perception algorithms. To develop efficient and scalable perception algorithms, the maximum information should be extracted from the available sensor data. In this work, we present our concept for an end-to-end perception architecture, named DeepSTEP. The deep learning-based architecture processes raw sensor data from the camera, LiDAR, and RaDAR, and combines the extracted data in a deep fusion network. The output of this deep fusion network is a shared feature space, which is used by perception head networks to fulfill several perception tasks, such as object detection or local mapping. DeepSTEP incorporates multiple ideas to advance state of the art: First, combining detection and localization into a single pipeline allows for efficient processing to reduce computational overhead and further improves overall performance. Second, the architecture leverages the temporal domain by using a self-attention mechanism that focuses on the most important features. We believe that our concept of DeepSTEP will advance the development of end-to-end perception systems. The network will be deployed on our research vehicle, which will be used as a platform for data collection, real-world testing, and validation. In conclusion, DeepSTEP represents a significant advancement in the field of perception for autonomous vehicles. The architecture’s end-to-end design, time-aware attention mechanism, and integration of multiple perception tasks make it a promising solution for real-world deployment. This research is a work in progress and presents the first concept of establishing a novel perception pipeline.

Autonomous racing with scaled race cars has gained increasing attention as an effective approach for developing perception, planning and control algorithms for safe autonomous driving at the limits of the vehicle’s handling. To train agile control policies for autonomous racing, learning-based approaches largely utilize reinforcement learning, albeit with mixed results. In this study, we benchmark a variety of imitation learning policies for racing vehicles that are applied directly or for bootstrapping reinforcement learning both in simulation and on scaled real-world environments. We show that interactive imitation learning techniques outperform traditional imitation learning methods and can greatly improve the performance of reinforcement learning policies by bootstrapping thanks to its better sample efficiency. Our benchmarks provide a foundation for future research on autonomous racing using Imitation Learning and Reinforcement Learning.

The driving behavior of autonomous vehicles has a significant impact on safety for all traffic participants. Unlike current traffic participants, autonomous vehicles in the future will also need to adhere to safety standards and defined risk properties in order to achieve a high level of public acceptance. At the same time, successful autonomous vehicles must be able to interact with human drivers in mixed traffic in a way that enables traffic to flow. In this paper, we present a hybrid approach to trajectory planning that learns and adapts human driving behavior using inverse reinforcement learning. The proposed approach performs a large-scale simulation with HighD real-world scenarios to learn human driving behavior and domain-specific traffic-flow characteristics. The analysis of the work focuses on the influence of risk-taking, which provides information about driving style safety. The results show insights into the risk behavior of trajectory planning approaches compared to human risk assessment. The comparison to human trajectories is intended to ensure comparability and accurate classification of risk-taking. We recommend a hybrid method for adapting driving behavior, in order to maintain the explainability and safety of the trajectory planning algorithm.

ROS 2 is a framework consisting of software libraries for developing robot systems, such as autonomous driving systems, that consist of multiple interacting components. In ROS 2, each component is implemented as a node, which contains time-triggered and event-triggered tasks. These tasks communicate with each other via ROS 2 topics or shared memory, and are scheduled by a ROS 2 executor. In ROS 2 systems, the system configuration and callback execution can have a significant impact on system performance, including end-to-end latencies, message loss, and memory usage. In this paper, we provide a bound on the timer period of ROS 2 timers to prevent sensor undersampling, and a subscription buffer size limit to prevent message loss and minimize memory usage. Furthermore, we explain the occurrence of message loss and high end-to-end latencies in ROS 2 systems, which are caused by the system configuration and subscription buffer size choice. Based on our observations, we propose a callback-prioritization heuristic to reduce end-to-end latencies and subscription buffer sizes. We demonstrate our findings using case studies based on Autoware.Universe and provide further evaluation to highlight the benefits of our heuristic.

High-performance autonomy often must operate at the boundaries of safety. When external agents are present in a system, the process of ensuring safety without sacrificing performance becomes extremely difficult. In this paper we present an approach to stress test such systems based on the rapidly exploring random tree (RRT) algorithm.We propose to find faults in such systems through adversarial agent perturbations, where the behaviors of other agents in an otherwise fixed scenario are modified. This creates a large search space of possibilities, which we explore both randomly and with a focused strategy that runs RRT in a bounded projection of the observable states that we call the objective space. The approach is applied to generate tests for evaluating overtaking logic and path planning algorithms in autonomous racing, where the vehicles are driving at high speed in an adversarial environment. We evaluate several autonomous racing path planners, finding numerous collisions during overtake maneuvers in all planners. The focused RRT search finds several times more crashes than the random strategy, and, for certain planners, tens to hundreds of times more crashes in the second half of the track.

Dampers are crucial components of the vehicle’s suspension to enable safe and comfortable driving. Therefore, defects like an oil leakage or a gas loss need to be detected expeditiously and with high accuracy. In this paper, we present a novel approach that relies solely on wheel speed signals to detect continuous levels of damper degradation. A dedicated 100000km real-world driving data set with multiple relevant damper defects and diverse environmental conditions is used for development and validation. Different vehicle types, routes, vehicle loads, tires, and driving styles are taken into account. Our approach comprises a frequency analysis of the wheel speed signals using the Fast Fourier Transform (FFT). A physical connection between defective dampers and oscillations in the wheel speeds enables a regression model to detect defective dampers. By using the residual sum of a polynomial fit of the FFT data points as a regressor variable, the current level of oil loss is determined. Subsequently, the remaining useful life (RUL) of the damper can be extrapolated. The resulting method is a threefold cascaded regression. In our results, we show a high sensitivity of the damper defect detection to vehicle loads as well as low sensitivity to ambient temperatures and rim sizes. The proposed method achieves a mean absolute error (MAE) of 5.4% oil loss. Future research will focus on efficiently implementing the algorithms onboard the vehicle and sending aggregated data to a remote back end for further analysis.

The 3rd Japan Automotive AI Challenge was an international online autonomous racing challenge where 164 teams competed in December 2021. This paper outlines the winning strategy to this competition, and the advantages and challenges of using the Autoware.Auto open source autonomous driving platform for multi-agent racing. Our winning approach includes a lane-switching opponent overtaking strategy, a global raceline optimization, and the integration of various tools from Autoware.Auto including a Model-Predictive Controller. We describe the use of perception, planning and control modules for high-speed racing applications and provide experience-based insights on working with Autoware.Auto. While our approach is a rule-based strategy that is suitable for non-interactive opponents, it provides a good reference and benchmark for learning-enabled approaches.

Scenario understanding and motion prediction are essential components for completely replacing human drivers and for enabling highly and fully automated driving (SAE-Level 4/5). In deeply stochastic and uncertain traffic scenarios, autonomous driving software must act beyond existing traffic rules and must predict critical situations in advance to provide safe and comfortable rides. In addition, comprehensive prediction models intend not just to reproduce, but rather to encode the human driver behavior, which requires profound scenario understanding. Hence, research in the field of scenario understanding and motion prediction also contributes to enable intelligent driver behavior models in general. This paper aims to review the state of research and outline common methods. A classification of these models is proposed according to their underlying investigation methodology. Based on this classification, a comparison is drawn between three specific prediction methods, which considers specific functional aspects and general requirements of applicability. The results of the comparison reveal a trade-off between holism and explainability in the state of the art. In conclusion, suggestions for future research objectives to solve this conflict are proposed.

The prediction of other road users is an essential task in autonomous driving for preventing collisions and enabling dynamic trajectory planning. This task becomes even more complex because different road users have different driving behaviors. There are underlying intentions that cannot be predicted with certainty without direct communication. In the current state of the art, most promising pattern-based models are trained on a dataset and then applied in the real world. In this paper we present an algorithm for vehicle trajectory prediction that is using online learning. The algorithm uses observations during the inference to optimize the underlying neural network at runtime. We show that our model can adapt to an observed behavior and thus improve the predicted uncertainty of trajectory predictions. Furthermore, we emphasize that our online learning approach can be transferred to many problems in self-supervised learning. The code used in this research is available as open-source software: https://github.com/TUMFTM/Wale-Net

We propose a prediction model to forecast the coverage of road networks in vehicle-to-vehicle or vehicle-to-infrastructure (V2X) networks for cloud-based automotive applications. The model is derived from fleet tests in the City of Munich (Germany). It considers the fleet and the road network characteristics by splitting the network into sub-networks and using the fleet's relative mileage on the sub-networks. The correlation of the spatial coverage and the fleet's mileage is analyzed for each sub-network showing that the expected degressive correlation exists. The derived regression model also shows a comparable fit for a data series taking the driving direction into account. Finally, we validated the model's ability to predict the temporal coverage by reducing the considered time intervals and taking the number of observations into account. The results show that the model can be used to predict the availability, the up-to-dateness and the accuracy of extended floating car data (XFCD).

This paper presents a detailed description of the software architecture that is used in the autonomous Roborace vehicles by the TUM-Team. The development of the software architecture was driven by both hardware components and usage of open source languages for making the software architecture reusable and easy to understand. The architecture combines the autonomous software functions perception, planning and control which are modularized for the usage on different hardware and for the purpose of using the car on high speed racetracks. The goal of the paper is to show which software functions are necessary for letting the car drive autonomously and fast around a racetrack.

The usage of battery electric vehicles in the commercial sector provides a lot of advantages. However many commercial company owners are reluctant to switch to battery electric vehicles. Important reasons for this are the insufficient range of battery electric vehicles and the lack of charging infrastructure. This paper presents an analysis of the charging infrastructure for battery electric vehicles in the commercial sector. The analysis is based on fleet test data, which was collected by 16 different commercial companies and 32 individual vehicles in the area of Munich. The approach in this paper is to use a simulation model, in which a backward-facing longitudinal dynamic model can simulate different types of electric vehicles. In addition, a charging simulation is integrated which includes the charging behavior of the user and different types of charging stations in the fleet test area of Munich. With the electric vehicle and charging simulation, it is possible to evaluate the current and future charging infrastructure of Munich. A comparison between public, company, employee and customer charging station locations is displayed in the results.

This paper presents an evaluation of the potential of integrating electric vehicles into commercial companies. The evaluation is based on fleet test data which was collected for 16 different commercial companies. The basic idea is to use a simulation, in which a backward-facing longitudinal dynamic model can simulate different types of electric vehicles. With this model and the real life velocity profiles from the fleet test data, the requested energy demand by the traction battery can be calculated. In addition, a charging simulation is integrated which includes the charging behavior of the user and the different types of charging stations in the fleet test area. With the electric vehicle and charging simulation it is possible to evaluate two questions: firstly, is it possible to substitute the companies' conventional cars with electric vehicles? Secondly, what influence does the current charging infrastructure has on the mobility of electric vehicles in the commercial sector?