This paper presents a sensorial-cooperative architecture to detect, track and classify entities in semi-structured outdoor scenarios for intelligent vehicles. In order to accomplish this task, information provided by in-vehicle Lidar and monocular vision is used. The detection and tracking phases are performed in the laser space, and the object classification methods work both in laser space (using a Gaussian Mixture Model classifier) and in vision spaces (AdaBoost classifier). A Bayesian-sum decision rule is used in order to combine the results of both classification techniques, and hence a more reliable object classification is achieved. Experiments confirm the effectiveness of the proposed architecture.

Object tracking is one of the key components of the perception system of autonomous cars and ADASs. With tracking, an ego-vehicle can make a prediction about the location of surrounding objects in the next time epoch and plan for next actions. Object tracking algorithms typically rely on sensory data (from RGB cameras or LIDAR). In fact, the integration of 2D-RGB camera images and 3D-LIDAR data can provide some distinct benefits. This paper proposes a 3D object tracking algorithm using a 3D-LIDAR, an RGB camera and INS (GPS/IMU) sensors data by analyzing sequential 2D-RGB, 3D point-cloud, and the ego-vehicle's localization data and outputs the trajectory of the tracked object, an estimation of its current velocity, and its predicted location in the 3D world coordinate system in the next time-step. Tracking starts with a known initial 3D bounding box for the object. Two parallel mean-shift algorithms are applied for object detection and localization in the 2D image and 3D point-cloud, followed by a robust 2D/3D Kalman filter based fusion and tracking. Reported results, from both quantitative and qualitative experiments using the KITTI database demonstrate the applicability and efficiency of the proposed approach in driving environments.

This paper addresses the problem of vehicle detection using Deep Convolutional Neural Network (ConvNet) and 3D-LIDAR data with application in advanced driver assistance systems and autonomous driving. A vehicle detection system based on the Hypothesis Generation (HG) and Verification (HV) paradigms is proposed. The data inputted to the system is a point cloud obtained from a 3D-LIDAR mounted on board an instrumented vehicle, which is transformed to a Dense-depth Map (DM). The proposed solution starts by removing ground points followed by point cloud segmentation. Then, segmented obstacles (object hypotheses) are projected onto the DM. Bounding boxes are fitted to the segmented objects as vehicle hypotheses (the HG step). Finally, the bounding boxes are used as inputs to a ConvNet to classify/verify the hypotheses of belonging to the category `vehicle' (the HV step). In this paper, we present an evaluation of ConvNet using LIDAR-based DMs and also the impact of domain-specific data augmentation on vehicle detection performance. To train and to evaluate the proposed vehicle detection system, the KITTI Benchmark Suite was used.

Robotic walkers are assistive robotic devices that provide mobility assistance, in a domestic or clinical scenario, to individuals suffering from a gait disorder, being age related or due to injuries, surgery, or diseases. Walkers also provide a significant potential for lower limb rehabilitation. In this paper, we present a novel multimodal robotic walker platform, the ISR-AIWALKER, where innovative contributions were made both in the human- machine interface (HMI) and in a gait analysis system placed on board the platform. Taking into account the application potential of these devices, an effort was made to use low-cost sensors without sacrificing the overall performance of the system. A change was made in the HMI paradigm, moving from a force-sensing to a vision-based approach, while maintaining a natural user interaction and adding complementary safety features like correct gripping enforcement. To cope with the close proximity of the user's body, a multimodal sensor setup was considered. Using both RGB and depth map data, a kinematic model of the user's lower limbs is obtained, allowing the identification of a set of features that are used in a machine learning approach to discriminate gait asymmetries. Experiments made with several subjects revealed that the proposed HMI is able to correctly estimate the user intention in a natural and intuitive way. The gait analysis system was also evaluated and evidenced a good discrimination capability to distinguish between different gait patterns.