Powered-Two-Wheelers (PTW) riders’ fatalities are prevalent on bends outside built-up areas due to the complexity and instability of their vehicles: countermeasures require a better understanding of the rider-PTW interaction. Analysing riding data is effective but becomes challenging when using extensive datasets; segmenting the riding data would help identify events of interest, isolate specific manoeuvres and describe the riding session. Manual segmentation would be time-consuming and subjective; automation would be beneficial. This work proposed an automatic, unsupervised tool for segmenting and clustering signals acquired during a riding session for studying motorcycle lateral dynamics in-depth. The method only requires measuring the motorcycle roll angle. An expert rider completed a closed route using an instrumented motorcycle; the algorithm divided the time series into segments categorised into clusters relative to specific riding conditions. Analysing the segmented trial revealed the effectiveness and usefulness of the approach. Then, a corner entry manoeuvre was investigated in-depth to observe each segment’s properties. The method associated each riding primitive to a cluster and described each manoeuvre through the segments’ succession. The clusters were unambiguous and easy to interpret thanks to their dynamics-based nature and minimal overlap. The algorithm identified the differences between the three corner entry manoeuvres in the trial. The segmentation simplified the in-depth corner entry analysis and allowed early detection of the manoeuvre start. The proposed tool can aid research on motorcycle dynamics, PTW-rider interaction, and riding preferences in bends. The segmented time series could be employed for rider training and pre-crash fall dynamics reconstruction.

Understanding driver-vehicle interactions remains a challenge, particularly in the case of cornering. This is particularly the case for powered two-wheeler vehicle (PTWs) users, perhaps because PTW drivers play a greater role in controlling the stability of their vehicles than do four-wheeled vehicle drivers. This difficulty stems from the variety of practices of this population of road users when entering, controlling their path, and exiting turns. Thus, observing the evolution of rider behavior during a cornering maneuver is an essential step in identifying road environment features that are risk factors for this category of road users. The real data set used for the experiments reported here was collected in the framework of the VIROLO++ collaborative project to improve the knowledge of the real practices of PTWs drivers, particularly during cornering. The in-depth analysis of these data in order to better understand motorcyclists’ behavior can therefore be considered a challenge. For this purpose, a two-step methodology was applied: (1) a data segmentation and feature extraction step in which the multidimensional time series of roll angle and roll velocity data were segmented using a multiple regression hidden logistic process (MRHLP), and (2) a clustering step in which the two detected segments characterizing the curve entry were assigned to different clusters regarding the curve initiation and the control actions set up during the different phases of the turning maneuver based on the hierarchical clustering algorithm. The results obtained show the effectiveness of the proposed methodology.

Indoor localization for mobile industrial robots is a crucial step toward an autonomous system. A mobile robot needs a reliable and robust localization system to achieve its task autonomously. A reasonable estimate of the robot's state can be achieved through Visual Odometry (VO); however, with dynamic objects in the scene, classical VO approaches need to detect and filter these moving objects. Alternatively, we can use an up-facing camera to track the movement with respect to the ceiling, which represents a static and invariant space. This paper presents Ceiling-DSO: an indoor ceiling-vision (CV) system based on Direct Sparse Odometry (DSO). We take advantage of the generic formulation of DSO to avoid making assumptions about the observable shapes or landmarks on the ceiling, making the method generic and applicable to multiple ceiling types. We built a ceiling-vision dataset in a real-world scenario; we then used it to test our approach with different DSO parameters to identify the best fit for robot pose estimation. This paper provides a qualitative and quantitative analysis of the obtained results that showed an acceptable error rate compared to the ground truth.

International audience

In this work, we investigate the accuracy of attitude estimation solutions based on inertial/magnetic sensors in the context of motorcycle trajectory reconstruction. We propose a comparative study and analysis of a number of quaternion based attitude-filtering methods. We conduct an experimental study using a multi-sensor platform with a precise ground truth obtained with a GPS RTK. We quantify the error in the motorcycle trajectory estimation obtained with each method and we discuss the obtained results.

This paper deals simultaneously with the trajectory estimation and map reconstruction by means of a stereo-calibrated vision system evolving in a large-scale unknown environment. This problem is widely known as Visual SLAM. Our proposal optimizes the execution time of the VSLAM framework while preserving its localization accuracy. The contributions of this paper are structured as follows. First, a novel VSLAM approach based on a “Weighted Mean” of multiple neighbor poses is detailed and is denoted as HOOFR SLAM. This approach provides a localization estimate after computing the camera poses (6-DOF rigid transformation) from the current image frame to previous neighbor frames. Taking advantage of the camera motion, we conjointly incorporate two types of stereo modes: “Static Stereo” mode (SS) through the fixed-baseline of left-right cameras setup along with the “Temporal Multi-view Stereo” mode (TMS). Moreover, instead of computing beforehand the disparity of SS mode for all key-points set, the disparity map in scale estimation step is limited to the inliers of the TMS mode so as to reduce the computational cost. This strategy is suitable to be parallelized on a multiprocessor architecture and exhibits a competitive performance with the other state-of-the-art strategies in many real datasets. Second, we report a hardware-software mapping of the proposed VSLAM approach. To this end, a heterogeneous CPU-GPU architecture-based vision system is considered. Third, a thorough and extensive experimental evaluation of our algorithm implemented on an automotive architecture (the NVIDIA Tegra TX1 system) is studied and analyzed. We report hence the localization and timing results through experiments on five well-known public stereo SLAM datasets: KITTI, Malaga, Oxford, MRT, and St_Lucia datasets.

The development of SLAM algorithms in the era of autonomous navigation and the growing demand for autonomous robot in place of human being, has put into question how to reduce the computational complexity and make use of these algorithms to operate in real time. Our work aims to take advantage of the high level synthesis (HLS) on FPGAs to design a real time SLAM application. Precisely, we evaluate the promess-held by the new modern low power FPGAs in accelerating SLAM algorithms. Throughtout this, we will attempt to implement a well-known algorithms (FastSLAM2.0), on a new modern FPGA using OpenCL, a standard high level language. Our implementation results show a significant improvement of the algorithm processing time on an FPGA device over a modern powerful embedded GPGPU.

In this paper, we propose an algorithm called Hessian ORB — Overlapped FREAK (HOOFR) to attack the feature matching problem in image processing. Our algorithm is based on the combination of the ORB detector and the FREAK bio-inspired descriptor. We address some modifications related to the detection and the description processes in order to enhance HOOFR reliability, speed and memory fingerprint. The experiments on a widely used dataset demonstrate the considerable performance of HOOFR compared to SIFT, SURF or ORB in terms of the execution time and the matching quality for various matching contexts. Moreover, the performance is also tested by integrating HOOFR in object-tracking application on ODROID-XU4 platform, presenting a high potential of its implementation on embedded systems.

International audience

Feature matching plays an important role in many computer vision applications, such as object recognition, scene reconstruction or image mosaicing. In this paper, we propose an algorithm called Hessian ORB - Overlapped FREAK (HOOFR) which is based on the combination of the ORB detector and the FREAK bio-inspired descriptor. We address some modifications related to the detection and the description processes in order to enhance HOOFR reliability, speed and memory fingerprint. The experiments on a widely used dataset demonstrate the considerable performance of HOOFR compared to SIFT, SURF or ORB in terms of the execution time and the matching quality, in various matching contexts.

In this paper we present a study of using an heterogeneous architecture to implement the graph-based SLAM algorithm. The study aims to investigate the performances of an ARM-GPU based architecture by offloading some critical compute-intensive tasks of the algorithm to the integrated GPU.

The graph-based SLAM (Simultaneous Localization and Mapping) method uses a graph to represent and solve the SLAM problem. The SLAM allows building a map of an unknown environment and simultaneously localizing the robot on this map. This paper presents a temporal analysis of the 3D graph-based SLAM method. We also propose an efficient implementation, on an OMAP embedded architecture, which is a widely used open multimedia applications platform. We provide an optimized data structure and an efficient memory access management to solve the nonlinear least squares problem related to the algorithm. The algorithm takes advantage of the Schur complement to reduce the execution time. We will present an optimized implementation of this task. We also take advantage of the multi-core architecture to parallelize the algorithm. To evaluate our implementation, we will compare the computational performances to the well known framework g2o. This work aims to demonstrate how optimizing data structure and multi-threading can decrease significantly the execution time of the graph-based SLAM on a low-cost architecture dedicated to embedded applications.

The first method that was developed to deal with the SLAM problem is based on the extended Kalman filter, EKF SLAM. However this approach cannot be applied to a large environments because of the quadratic complexity and data association problem. The second approach to address the SLAM problem is based on the Rao-Blackwellized Particle filter FastSLAM, which follows a large number of hypotheses that represent the different possible trajectories, each trajectory carries its own map, its complexity increase logarithmically with the number of landmarks in the map. In this paper we will present the result of an implementation of the FastSLAM 2.0 on an open multimedia applications processor, based on a monocular camera as an exteroceptive sensor. A parallel implementation of this algorithm was achieved. Results aim to demonstrate that an optimized algorithm implemented on a low cost architecture is suitable to design an embedded system for SLAM applications.

An autonomous robot has to localize itself in an unknown area. Simultaneous Localization and Mapping (SLAM) allows for a robot to build a map of an unknown environment and localize simultaneously itself on this map. Graph-based SLAM methods use a graph to represent and solve the SLAM problem. This paper presents an optimized implementation of the incremental 3D graph-based SLAM on an OMAP architecture used as open multimedia applications platform. This implementation uses an optimized data structure and an efficient memory access management to solve the nonlinear least squares problem related to the algorithm. It takes also advantage of the multi-core architecture to parallelize the algorithm. To evaluate our implementation, we will evaluate the processing times of the implemented algorithm compared to those of the well known framework g2o.

This paper presents the design process of a desk-set tangible user interface for the navigation and manipulation of media content organized by content-based similarity with off-the-shelf/flea market devices. For intra-media navigation, a refurbished portable vinyl player has its inside mechanics replaced by a webcam monitoring circular gray code analyzed through computer vision for position/speed tracking. For inter-media navigation, a 3D force-feedback controller is mounted in upright position on a truss with cell clamps, repurposed as trackpad. For media recomposition, motorized faders recall the effect presets of the closest/last selected media item.

This paper describes a multi-FPGA architecture dedicated for the real-time imaging using the Total Focusing Method (TFM) and an advanced acquisition technique called the Full Matrix Capture (FMC). The maximum operating frequency is 147.3 MHz for the control blocks and 161.3 MHz for the acquisition blocks. The architecture is able to perform real-time image reconstruction at a maximum frame rate of 73 fps and a maximum resolution of 128 × 128 pixels, which is adequate as a performance for a real-time imaging system with a good characterization of defects.