An open problem in autonomous driving research is modeling human driving behavior, which is needed for the planning component of the autonomy stack, safety validation through traffic simulation (TS), and causal inference for generating explanations for autonomous driving. Modeling human driving behavior is challenging because it is stochastic, high-dimensional, and involves interaction between multiple agents. This problem has been studied in various communities with a vast body of the literature. Existing reviews have generally focused on one aspect: motion prediction (MP). In this article, we present a unification of the literature that covers intent estimation, trait estimation (TE), and motion prediction. This unification is enabled by modeling multiagent driving as a partially observable stochastic game (POSG), which allows us to cast driver modeling tasks as inference problems. We classify driver models into a taxonomy based on the specific tasks they address and the key attributes of their approach. Finally, we identify open research opportunities in the field of driver modeling.

Hand gesture recognition (HGR) plays an essential role in human–computer interaction, enabling natural, touch-free control across domains such as virtual reality, sign language interpretation, and automotive interfaces. Despite notable progress using deep networks and multimodal data fusion, current HGR solutions still face challenges stemming from misclassifications and noise sensitivity. In this paper, we propose HomE, a homogeneous ensemble framework that aims to improve HGR models’ performance and robustness by partitioning gesture classes into smaller, more coherent subsets based on critical features uncovered in parallel by unsupervised clustering and an LLM-driven semantic sampler, and by training a dedicated expert learner for each subset. Then, a separate router learner routes incoming samples to the most relevant expert learner, while the expert routing module fuses the outputs of all expert learners into a final classification. Extensive experiments on the NVGestures, DHG-14, and SHREC’17 datasets show that our method not only enhances accuracy and robustness over single-network baselines but also enables these base models to become more competitive with state-of-the-art approaches—all without altering their underlying architectures. Furthermore, our ablation studies verify that multiple heterogeneous sampling methods provide complementary strengths, ultimately boosting recognition performance. In addition to offering insights on sampling strategies, this work highlights the scalability of HomE for both depth and skeleton-based HGR tasks, suggesting its broader applicability to other domains where class diversity and label ambiguity pose obstacles for single-model approaches.

Teleoperation is an important technology to enable supervisors to control agricultural robots remotely. However, environmental factors in dense crop rows and limitations in network infrastructure hinder the reliability of data streamed to teleoperators. These issues result in delayed and variable frame rate video feeds that often deviate significantly from the robot's actual viewpoint. We propose a modular learning-based vision pipeline to generate delay-compensated images in real-time for supervisors. Our extensive offline evaluations demonstrate that our method generates more accurate images compared to state-of-the-art approaches in our setting. Additionally, ours is one of the few works to evaluate a delay-compensation method in outdoor field environments with complex terrain on data from a real robot in real-time. Resulting videos and code are provided at https://sites.google.com/illinois.edu/comp-teleop.

When performing tasks like laundry, humans naturally coordinate both hands to manipulate objects and anticipate how their actions will change the state of the clothes. However, achieving such coordination in robotics remains challenging due to the need to model object movement, predict future states, and generate precise bimanual actions. In this work, we address these challenges by infusing the predictive nature of human manipulation strategies into robot imitation learning. Specifically, we disentangle task-related state transitions from agent-specific inverse dynamics modeling to enable effective bimanual coordination. Using a demonstration dataset, we train a diffusion model to predict future states given historical observations, envisioning how the scene evolves. Then, we use an inverse dynamics model to compute robot actions that achieve the predicted states. Our key insight is that modeling object movement can help learning policies for bimanual coordination manipulation tasks. Evaluating our framework across diverse simulation and real-world manipulation setups, including multimodal goal configurations, bimanual manipulation, deformable objects, and multi-object setups, we find that it consistently outperforms state-of-the-art state-to-action mapping policies. Our method demonstrates a remarkable capacity to navigate multimodal goal configurations and action distributions, maintain stability across different control modes, and synthesize a broader range of behaviors than those present in the demonstration dataset.

Employing a teleoperation system for gathering demonstrations offers the potential for more efficient learning of robot manipulation. However, teleoperating a robot arm equipped with a dexterous hand or gripper, via a teleoperation system presents inherent challenges due to the task's high dimensionality, complexity of motion, and differences between physiological structures. In this study, we introduce a novel system for joint learning between human operators and robots, that enables human operators to share control of a robot end-effector with a learned assistive agent, simplifies the data collection process, and facilitates simultaneous human demonstration collection and robot manipulation training. As data accumulates, the assistive agent gradually learns. Consequently, less human effort and attention are required, enhancing the efficiency of the data collection process. It also allows the human operator to adjust the control ratio to achieve a tradeoff between manual and automated control. We conducted experiments in both simulated environments and physical realworld settings. Through user studies and quantitative evaluations, it is evident that the proposed system could enhance data collection efficiency and reduce the need for human adaptation while ensuring the collected data is of sufficient quality for downstream tasks. For more details, please refer to our webpage https://norweig1an.github.io/HAJL.github.io/.

Using formal methods to evaluate software and hardware enhances system reliability, which is crucial for safety-critical applications such as airplanes and autonomous vehicles. Formal methods are mathematical modeling techniques that can be used to verify the safety of systems. The use of formal methods is limited in industry due to a shortage of trained engineers. Educators in formal methods often report that many students do not see the benefit of formal methods and perceive the involved math as not worth the effort for their future careers as software engineers. This study aims to understand the current state of student beliefs and how using a formal verification tool affects student motivation to learn about formal methods. We used an Expectancy Value Cost Lite survey to measure student motivation. Students completed this survey multiple times while designing algorithms to control vehicles in different scenarios, both with and without a formal verification tool. We found that students in an autonomy class are motivated to use formal methods. Although the findings are not statistically significant, we observed a slight increase in motivation after using the tool. Additionally, using a formal verification tool solely for modeling may contribute to increased motivation. These results suggest that incorporating tools into coursework may be a useful step in motivating more students to study formal methods and enter the workforce with these skills.

Sampling-based planning algorithms like Rapidly-exploring Random Tree (RRT) are versatile in solving path planning problems. RRT* offers asymptotic optimality but requires growing the tree uniformly over the free space, which leaves room for efficiency improvement. To accelerate convergence, rule-based informed approaches sample states in an admissible ellipsoidal subset of the space determined by the current path cost. Learning-based alternatives model the topology of the free space and infer the states close to the optimal path to guide planning. We propose Neural Informed RRT* to combine the strengths from both sides. We define point cloud representations of free states. We perform Neural Focus, which constrains the point cloud within the admissible ellipsoidal subset from Informed RRT*, and feeds into PointNet++ for refined guidance state inference. In addition, we introduce Neural Connect to build connectivity of the guidance state set and further boost performance in challenging planning problems. Our method surpasses previous works in path planning benchmarks while preserving probabilistic completeness and asymptotic optimality. We deploy our method on a mobile robot and demonstrate real world navigation around static obstacles and dynamic humans. Code is available at https://github.com/tedhuang96/nirrt_star.

Based on the recent advancements in representation learning, we propose a novel pipeline for task-oriented voice-controlled robots with raw sensor inputs. Previous methods rely on a large number of labels and task-specific reward functions. Not only can such an approach hardly be improved after the deployment, but also has limited generalization across robotic platforms and tasks. To address these problems, our pipeline first learns a visual-audio representation (VAR) that associates images and sound commands. Then the robot learns to fulfill the sound command via reinforcement learning using the reward generated by the VAR. We demonstrate our approach with various sound types, robots, and tasks. We show that our method outperforms previous work with much fewer labels. We show in both the simulated and real-world experiments that the system can self-improve in previously unseen scenarios given a reasonable number of newly labeled data.

Collaborative robots require effective human intention estimation to safely and smoothly work with humans in less structured tasks such as industrial assembly, where human intention continuously changes. We propose the concept of intention tracking and introduce a collaborative robot system that concurrently tracks intentions at hierarchical levels. The high-level intention is tracked to estimate human's interaction pattern and enable robot to (1) avoid collision with human to minimize interruption and (2) assist human to correct failure. The low-level intention estimate provides robot with task-related information. We implement the system on a UR5e robot and demonstrate robust, seamless and ergonomic human-robot collaboration in an ablative pilot study of an assembly use case.

We study the problem of safe and intention-aware robot navigation in dense and interactive crowds. Most previous reinforcement learning (RL) based methods fail to consider different types of interactions among all agents or ignore the intentions of people, which results in performance degradation. In this paper, we propose a novel recurrent graph neural network with attention mechanisms to capture heterogeneous interactions among agents through space and time. To encourage longsighted robot behaviors, we infer the intentions of dynamic agents by predicting their future trajectories for several timesteps. The predictions are incorporated into a model-free RL framework to prevent the robot from intruding into the intended paths of other agents. We demonstrate that our method enables the robot to achieve good navigation performance and non-invasiveness in challenging crowd navigation scenarios. We successfully transfer the policy learned in simulation to a real-world TurtleBot 2i. Our code and videos are available at https://sites.google.com/view/intention-aware-crowdnav/home.

Autonomous navigation in crowded spaces poses a challenge for mobile robots due to the highly dynamic, partially observable environment. Occlusions are highly prevalent in such settings due to a limited sensor field of view and obstructing human agents. Previous work has shown that observed interactive behaviors of human agents can be used to estimate potential obstacles despite occlusions. We propose integrating such social inference techniques into the planning pipeline. We use a variational autoencoder with a specially designed loss function to learn representations that are meaningful for occlusion inference. This work adopts a deep reinforcement learning approach to incorporate the learned representation into occlusion-aware planning. In simulation, our occlusion-aware policy achieves comparable collision avoidance performance to fully observable navigation by estimating agents in occluded spaces. We demonstrate successful policy transfer from simulation to the real-world Turtlebot 2i. To the best of our knowledge, this work is the first to use social occlusion inference for crowd navigation. Our implementation is available at https://github.com/yejimun/PaS_CrowdNav.

When humans interact with robots influence is inevitable. Consider an autonomous car driving near a human: the speed and steering of the autonomous car will affect how the human drives. Prior works have developed frameworks that enable robots to influence humans towards desired behaviors. But while these approaches are effective in the short-term (i.e., the first few human-robot interactions), here we explore long-term influence (i.e., repeated interactions between the same human and robot). Our central insight is that humans are dynamic: people adapt to robots, and behaviors which are influential now may fall short once the human learns to anticipate the robot's actions. With this insight, we experimentally demonstrate that a prevalent game-theoretic formalism for generating influential robot behaviors becomes less effective over repeated interactions. Next, we propose three modifications to Stackelberg games that make the robot's policy both influential and unpredictable. We finally test these modifications across simulations and user studies: our results suggest that robots which purposely make their actions harder to anticipate are better able to maintain influence over long-term interaction. See videos here: https://youtu.be/ydO83cgjZ2Q

The use of mobile robots in unstructured environments like the agricultural field is becoming increasingly common. The ability for such field robots to proactively identify and avoid failures is thus crucial for ensuring efficiency and avoiding damage. However, the cluttered field environment introduces various sources of noise (such as sensor occlusions) that make proactive anomaly detection difficult. Existing approaches can show poor performance in sensor occlusion scenarios as they typically do not explicitly model occlusions and only leverage current sensory inputs. In this work, we present an attention-based recurrent neural network architecture for proactive anomaly detection that fuses current sensory inputs and planned control actions with a latent representation of prior robot state. We enhance our model with an explicitly-learned model of sensor occlusion that is used to modulate the use of our latent representation of prior robot state. Our method shows improved anomaly detection performance and enables mobile field robots to display increased resilience to predicting false positives regarding navigation failure during periods of sensor occlusion, particularly in cases where all sensors are briefly occluded. Our code is available at: https://github.com/andreschreiber/roar.

Efficient transfer learning algorithms are key to the success of foundation models on diverse downstream tasks even with limited data. Recent works of Basu et al. (2023) and Kaba et al. (2022) propose group averaging (equitune) and optimization-based methods, respectively, over features from group-transformed inputs to obtain equivariant outputs from non-equivariant neural networks. While Kaba et al. (2022) are only concerned with training from scratch, we find that equitune performs poorly on equivariant zero-shot tasks despite good finetuning results. We hypothesize that this is because pretrained models provide better quality features for certain transformations than others and simply averaging them is deleterious. Hence, we propose λ-equitune that averages the features using importance weights, λs. These weights are learned directly from the data using a small neural network, leading to excellent zero-shot and finetuned results that outperform equitune. Further, we prove that λ-equitune is equivariant and a universal approximator of equivariant functions. Additionally, we show that the method of Kaba et al. (2022) used with appropriate loss functions, which we call equizero, also gives excellent zero-shot and finetuned performance. Both equitune and equizero are special cases of λ- equitune. To show the simplicity and generality of our method, we validate on a wide range of diverse applications and models such as 1) image classification using CLIP, 2) deep Q-learning, 3) fairness in natural language generation (NLG), 4) compositional generalization in languages, and 5) image classification using pretrained CNNs such as Resnet and Alexnet.

Spatio-temporal graphs (ST-graphs) have been used to model time series tasks such as traffic forecasting, human motion modeling, and action recognition. The high-level structure and corresponding features from ST-graphs have led to improved performance over traditional architectures. However, current methods tend to be limited by simple features, despite the rich information provided by the full graph structure, which leads to inefficiencies and suboptimal performance in downstream tasks. We propose the use of features derived from meta-paths, walks across different types of edges, in ST-graphs to improve the performance of Structural Recurrent Neural Network. In this paper, we present the Meta-path Enhanced Structural Recurrent Neural Network (MESRNN), a generic framework that can be applied to any spatio-temporal task in a simple and scalable manner. We employ MESRNN for pedestrian trajectory prediction, utilizing these meta-path based features to capture the relationships between the trajectories of pedestrians at different points in time and space. We compare our MESRNN against state-of-the-art ST-graph methods on standard datasets to show the performance boost provided by meta-path information. The proposed model consistently outperforms the baselines in trajectory prediction over long time horizons by over 32%, and produces more socially compliant trajectories in dense crowds. For more information please refer to the project website at https://sites.google.com/illinois.edu/mesrnn/home.

Autonomous vehicles must reason about spatial occlusions in urban environments to ensure safety without being overly cautious. Prior work explored occlusion inference from observed social behaviors of road agents, hence treating people as sensors. Inferring occupancy from agent behaviors is an inherently multimodal problem; a driver may behave similarly for different occupancy patterns ahead of them (e.g., a driver may move at constant speed in traffic or on an open road). Past work, however, does not account for this multimodality, thus neglecting to model this source of aleatoric uncertainty in the relationship between driver behaviors and their environment. We propose an occlusion inference method that characterizes observed behaviors of human agents as sensor measurements, and fuses them with those from a standard sensor suite. To capture the aleatoric uncertainty, we train a conditional variational autoencoder with a discrete latent space to learn a multimodal mapping from observed driver trajectories to an occupancy grid representation of the view ahead of the driver. Our method handles multi-agent scenarios, combining measurements from multiple observed drivers using evidential theory to solve the sensor fusion problem. Our approach is validated on a cluttered, real-world intersection, outperforming baselines and demonstrating real-time capable performance. Our code is available at https://github.com/sisl/MultiAgentVariationalOcclusionInferenc

Applying reinforcement learning (RL) methods on robots typically involves training a policy in simulation and deploying it on a robot in the real world. Because of the model mismatch between the real world and the simulator, RL agents deployed in this manner tend to perform suboptimally. To tackle this problem, researchers have developed robust policy learning algorithms that rely on synthetic noise disturbances. However, such methods do not guarantee performance in the target environment. We propose a convex risk minimization algorithm to estimate the model mismatch between the simulator and the target domain using trajectory data from both environments. We show that this estimator can be used along with the simulator to evaluate performance of an RL agents in the target domain, effectively bridging the gap between these two environments. We also show that the convergence rate of our estimator to be of the order of$n^{-1/4}$, where$n$is the number of training samples. In simulation, we demonstrate how our method effectively approximates and evaluates performance on Gridworld, Cartpole, and Reacher environments on a range of policies. We also show that the our method is able to estimate performance of a 7 DOF robotic arm using the simulator and remotely collected data from the robot in the real world.

Navigation through uncontrolled intersections is one of the key challenges for autonomous vehicles. Identifying the subtle differences in hidden traits of other drivers can bring significant benefits when navigating in such environments. We propose an unsupervised method for inferring driver traits such as driving styles from observed vehicle trajectories. We use a variational autoencoder with recurrent neural networks to learn a latent representation of traits without any ground truth trait labels. Then, we use this trait representation to learn a policy for an autonomous vehicle to navigate through a T-intersection with deep reinforcement learning. Our pipeline enables the autonomous vehicle to adjust its actions when dealing with drivers of different traits to ensure safety and efficiency. Our method demonstrates promising performance and outperforms state-of-the-art baselines in the T-intersection scenario.

We investigate how employing model learning methods in concert with model predictive control (MPC) can be used to automate obstacle reduction to mitigate risks to Combat Engineers operating construction equipment in an active battlefield. We focus on the task of earthen berm removal using a bladed vehicle. We introduce a novel data-driven formulation for earthmoving dynamics that enables prediction of the vehicle and detailed terrain state over a one second horizon. In a simulation environment, we first record demonstrations from a human operator and then train two different earthmoving models to produce predictions of the high-dimensional state using under six minutes of data. Optimization over the learned model is performed to select an action sequence, constrained to a 2D space of template action trajectories. Simple recovery controllers are implemented to improve controller performance when the model predictions degrade. This system yields near human-level performance on a berm removal task, indicating that model learning and predictive control is a promising data-efficient approach to autonomous earthmoving.

Agriculture is facing a labor crisis, leading to increased interest in fleets of small, under-canopy robots (agbots) that can perform precise, targeted actions (e.g., crop scouting, weeding, fertilization), while being supervised by human operators remotely. However, farmers are not necessarily experts in robotics technology and will not adopt technologies that add to their workload or do not provide an immediate payoff. In this work, we explore methods for communication between a remote human operator and multiple agbots and examine the impact of audio communication on the operator’s preferences and productivity. We develop a simulation platform where agbots are deployed across a field, randomly encounter failures, and call for help from the operator. As the agbots report errors, various audio communication mechanisms are tested to convey which robot failed and what type of failure occurs. The human is tasked with verbally diagnosing the failure while completing a secondary task. A user study was conducted to test three audio communication methods: earcons, single-phrase commands, and full sentence communication. Each participant completed a survey to determine their preferences and each method’s overall effectiveness. Our results suggest that the system using single phrases is the most positively perceived by participants and may allow for the human to complete the secondary task more efficiently. The code is available at: https://github.com/akamboj2/Agbot-Sim.

Uncovering potential failure cases is a crucial step in the validation of safety critical systems such as autonomous vehicles. Failure search may be done through logging substantial vehicle miles in either simulation or real world testing. Due to the sparsity of failure events, naive random search approaches require significant amounts of vehicle operation hours to find potential system weaknesses. As a result, adaptive searching techniques have been proposed to efficiently explore and uncover failure trajectories of an autonomous policy in simulation. Adaptive Stress Testing (AST) is one such method that poses the problem of failure search as a Markov decision process and uses reinforcement learning techniques to find high probability failures. However, this formulation requires a probability model for the actions of all agents in the environment. In systems where the environment actions are discrete and dependencies among agents exist, it may be infeasible to fully characterize the distribution or find a suitable proxy. This work proposes the use of a data driven approach to learn a suitable classifier that tries to model how humans identify critical states and use this to guide failure search in AST. We show that the incorporation of critical states into the AST framework generates failure scenarios with increased safety violations in an autonomous driving policy with a discrete action space.

Safe and efficient navigation through human crowds is an essential capability for mobile robots. Previous work on robot crowd navigation assumes that the dynamics of all agents are known and well-defined. In addition, the performance of previous methods deteriorates in partially observable environments and environments with dense crowds. To tackle these problems, we propose decentralized structural-Recurrent Neural Network (DS-RNN), a novel network that reasons about spatial and temporal relationships for robot decision making in crowd navigation. We train our network with model-free deep reinforcement learning without any expert supervision. We demonstrate that our model outperforms previous methods in challenging crowd navigation scenarios. We successfully transfer the policy learned in the simulator to a real-world TurtleBot 2i.

We explore the interpretation of sound for robot decision making, inspired by human speech comprehension. While previous methods separate sound processing unit and robot controller, we propose an end-to-end deep neural network which directly interprets sound commands for visual-based decision making. The network is trained using reinforcement learning with auxiliary losses on the sight and sound networks. We demonstrate our approach on two robots, a TurtleBot3 and a Kuka-IIWA arm, which hear a command word, identify the associated target object, and perform precise control to reach the target. For both robots, we show the effectiveness of our network in generalization to sound types and robotic tasks empirically. We successfully transfer the policy learned in simulator to a real-world TurtleBot3.

Recent developments in multi-agent imitation learning have shown promising results for modeling the behavior of human drivers. However, it is challenging to capture emergent traffic behaviors that are observed in real-world datasets. Such behaviors arise due to the many local interactions between agents that are not commonly accounted for in imitation learning. This paper proposes Reward Augmented Imitation Learning (RAIL), which integrates reward augmentation into the multi-agent imitation learning framework and allows the designer to specify prior knowledge in a principled fashion. We prove that convergence guarantees for the imitation learning process are preserved under the application of reward augmentation. This method is validated in a driving scenario, where an entire traffic scene is controlled by driving policies learned using our proposed algorithm. Further, we demonstrate improved performance in comparison to traditional imitation learning algorithms both in terms of the local actions of a single agent and the behavior of emergent properties in complex, multi-agent settings.

Imitation learning has proven to be useful for many real-world problems, but approaches such as behavioral cloning suffer from data mismatch and compounding error issues. One attempt to address these limitations is the DAgger algorithm, which uses the state distribution induced by the novice to sample corrective actions from the expert. Such sampling schemes, however, require the expert to provide action labels without being fully in control of the system. This can decrease safety and, when using humans as experts, is likely to degrade the quality of the collected labels due to perceived actuator lag. In this work, we propose HG-DAgger, a variant of DAgger that is more suitable for interactive imitation learning from human experts in real-world systems. In addition to training a novice policy, HG-DAgger also learns a safety threshold for a model-uncertainty-based risk metric that can be used to predict the performance of the fully trained novice in different regions of the state space. We evaluate our method on both a simulated and real-world autonomous driving task, and demonstrate improved performance over both DAgger and behavioral cloning.

Gradient-based methods are often used for policy optimization in deep reinforcement learning, despite being vulnerable to local optima and saddle points. Although gradient-free methods (e.g., genetic algorithms or evolution strategies) help mitigate these issues, poor initialization and local optima are still concerns in highly nonconvex spaces. This paper presents a method for policy optimization based on Monte-Carlo tree search and gradient-free optimization. Our method, called Monte-Carlo tree search for policy optimization (MCTSPO), provides a better exploration-exploitation trade-off through the use of the upper confidence bound heuristic. We demonstrate improved performance on reinforcement learning tasks with deceptive or sparse reward functions compared to popular gradient-based and deep genetic algorithm baselines.

Although imitation learning is often used in robotics, the approach frequently suffers from data mismatch and compounding errors. DAgger is an iterative algorithm that addresses these issues by aggregating training data from both the expert and novice policies, but does not consider the impact of safety. We present a probabilistic extension to DAgger, which attempts to quantity the confidence of the novice policy as a proxy for safety. Our method, EnsembleDAgger, approximates a Gaussian Process using an ensemble of neural networks. Using the variance as a measure of confidence, we compute a decision rule that captures how much we doubt the novice, thus determining when it is safe to allow the novice to act. With this approach, we aim to maximize the novice's share of actions, while constraining the probability of failure. We demonstrate improved safety and learning performance compared to other DAgger variants and classic imitation learning on an inverted pendulum and in the MuJoCo HalfCheetah environment.

Despite growing attention in autonomy, there are still many open problems, including how autonomous vehicles will interact and communicate with other agents, such as human drivers and pedestrians. Unlike most approaches that focus on pedestrian detection and planning for collision avoidance, this paper considers modeling the interaction between human drivers and pedestrians and how it might influence map estimation, as a proxy for detection. We take a mapping inspired approach and incorporate people as sensors into mapping frameworks. By taking advantage of other agents' actions, we demonstrate how we can impute portions of the map that would otherwise be occluded. We evaluate our framework in human driving experiments and on real-world data, using occupancy grids and landmark-based mapping approaches. Our approach significantly improves overall environment awareness and outperforms standard mapping techniques.

To improve efficiency and reduce failures in autonomous vehicles, research has focused on developing robust and safe learning methods that take into account disturbances in the environment. Existing literature in robust reinforcement learning poses the learning problem as a two player game between the autonomous system and disturbances. This paper examines two different algorithms to solve the game, Robust Adversarial Reinforcement Learning and Neural Fictitious Self Play, and compares performance on an autonomous driving scenario. We extend the game formulation to a semi-competitive setting and demonstrate that the resulting adversary better captures meaningful disturbances that lead to better overall performance. The resulting robust policy exhibits improved driving efficiency while effectively reducing collision rates compared to baseline control policies produced by traditional reinforcement learning methods.

Given the current capabilities of autonomous vehicles, one can easily imagine autonomous vehicles being released on the road in the near future. However, it can be assumed that this transition will not be instantaneous, suggesting that autonomous vehicles will have to be capable of driving in a mixed environment, with both humans and autonomous vehicles. To guarantee smooth integration and maintain the nuanced social interactions on the road, a shared mental model must be developed. This means that the behaviors of human-driven vehicles and their typical interactions in collaborative maneuvers must be modeled and understood in an accurate and precise manner. Then, by integrating such models into autonomous planning, we can develop control frameworks that mimic this shared understanding. We present a driver modeling framework that estimates an empirical reachable set to capture typical lane changing behaviors. This method can predict driver behaviors with up to 90% accuracy and cumulative errors less than 1 m. Leveraging this driver model in an optimization-based trajectory planning framework, we can generate trajectories that are similar to those performed by humans. By using this modeling and planning framework, we can improve understanding and integration of nuanced interactions to improve collaboration between humans and autonomy.

Given the current capabilities of autonomous vehicles, one can easily imagine autonomy released on the road in the near future. However, it can be assumed that the transition will not be instantaneous, meaning they will have to be capable of driving well in a mixed environment, with both humans and other autonomous vehicles on the road. This leaves a number of concerns for autonomous vehicles in terms of dealing with human uncertainty and understanding of cooperation on the road. This work demonstrates the need for focusing on communication and collaboration between autonomy and human drivers. After analyzing how drivers perform cooperative maneuvers (e.g. lane changing), key cues were identified for conveying intent through nonverbal communication. It was found that human observers can predict lane changes with over two seconds in prior to the lane departure, without use of a turning signal. Building on this concept, an autonomous control scheme is proposed that aims to capture these subtle motions before executing a lane change. To compare the proposed human-inspired methods, three possible control schemes for autonomous vehicles are implemented for a validation study on human subjects to provide feedback on their experience. By properly conveying intent through nuanced trajectory planning, we show that drivers can predict the autonomous vehicle's actions with 40% increase in prediction time when compared to traditional control methods, both as a passenger and while observing the autonomous vehicle.

In light of growing attention of intelligent vehicle systems, we have present an assessment of methods for driver models that predict driver behaviors. This work looks at varying datasets to see the affects on intent detection algorithms. The motivation is to understand and assess how data is mapped from datasets to discrete states or modes of intent. Using a model of a human driver's decision making process to estimate intent, we build techniques for analyzing and learning human behaviors to improve understanding. We derive models based off of human perception and interaction with the environment (e.g. other vehicles on the road), that is generalizable and flexible enough to detect intent across different drivers. The resulting detection scheme is able to determine driver intent with high accuracy across multiple drivers, relying on a large dataset consisting of lane changes under varying environmental constraints. By comparing different labeling methods, we assess the effectiveness of learned models under different class variations. This allows us to derive accurate and general models for detecting intent that rely on the subtle variations and behaviors that humans exhibit while driving.

In order to develop provably safe human-in-the-loop systems, accurate and precise models of human behavior must be developed. Driving is a good example of such a system because the driver has full control of the vehicle, and her likely actions are highly dependent on her mental state and the context of the current situation. This paper presents a testbed for collecting driver data that allows us to collect realistic data, while maintaining safety and control of the environmental surroundings. We extend previous work that focuses on set predictions consisting of trajectories observed from the nonlinear dynamics and behaviors of the human driven car, accounting for the driver mental state, the context or situation that the vehicle is in, and the surrounding environment in both highway and intersection scenarios. This allows us to predict driving behavior over long time horizons with extremely high accuracy. By using this realistic data and flexible algorithm, a precise and accurate driver model can be developed that is tailored to an individual and usable in semi-autonomous frameworks.

Existing commercial driver assistance systems, including automatic braking systems and lane-keeping systems, may monitor the state of the vehicle or the environment to determine whether the systems should intervene. However, the state of the human driver is not typically included in the decision making process. In this paper, we propose to use hidden mode stochastic hybrid systems to model the interaction between the human driver and the vehicle. We show that by monitoring the human behavior as well as the vehicle state, we can infer the human state and enhance the quality of decision making in a driver assistance system. The resulting control policy is obtained by solving an optimal planning problem of the proposed hidden mode hybrid system. The policy can automatically balance the decision making about when to give warning to the driver and when to actually intervene in the control of the vehicle.

Threat assessment during semiautonomous driving is used to determine when correcting a driver's input is required. Since current semiautonomous systems perform threat assessment by predicting a vehicle's future state while treating the driver's input as a disturbance, autonomous controller intervention is limited to a restricted regime. Improving vehicle safety demands threat assessment that occurs over longer prediction horizons wherein a driver cannot be treated as a malicious agent. In this paper, we describe a real-time semiautonomous system that utilizes empirical observations of a driver's pose to inform an autonomous controller that corrects a driver's input when possible in a safe manner. We measure the performance of our system using several metrics that evaluate the informativeness of the prediction and the utility of the intervention procedure. A multisubject driving experiment illustrates the usefulness, with respect to these metrics, of incorporating the driver's pose while designing a semiautonomous system.