Stephen Decker

Collision Sensing by Stereo Vision and Radar Sensor Fusion

To take the advantages of both stereo cameras and radar, this paper proposes a fusion approach to accurately estimate the location, size, pose and motion information of a threat vehicle with respect to the host from observations obtained by both sensors. To do that, we first fit the contour of a threat vehicle from stereo depth information, and find the closest point on the contour from the vision sensor. Then the fused closest point is obtained by fusing radar observations and the vision closest point. Next by translating the fitted contour to the fused closest point, the fused contour is obtained. Finally the fused contour is tracked by using the rigid body constraints to estimate the location, size, pose and motion of the threat vehicle. Experimental results from both the synthetic data and the real world road test data demonstrate the success of the proposed algorithm.

Collision sensing by stereo vision and radar sensor fusion

To take the advantages of both stereo cameras and radar, this paper proposes a fusion approach to accurately estimate the location, size, pose and motion information of a threat vehicle with respect to the host from observations obtained by both sensors. To do that, we first fit the contour of a threat vehicle from stereo depth information, and find the closest point on the contour from the vision sensor. Then the fused closest point is obtained by fusing radar observations and the vision closest point. Next by translating the fitted contour to the fused closest point, the fused contour is obtained. Finally the fused contour is tracked by using the rigid body constraints to estimate the location, size, pose and motion of the threat vehicle. Experimental results from both the synthetic data and the real world road test data demonstrate the success of the proposed algorithm.

A Simple Vehicle Model for Path Prediction During Evasive Maneuvers and a Stochastic Analysis on the Crash Probability

Active safety systems are being developed in automotive industry, and an analytical vehicle model is needed in such systems to predict vehicle path to assess the crash probability. However, the bicycle model cannot result in a good correlation with test data and ADAMS simulation results, and other analytical vehicle models which have 8 or 14 degrees of freedom need more computation time. Therefore, in this study a simple analytical vehicle model was proposed to predict vehicle path especially during evasive maneuvers. The analytical vehicle model can predict a vehicle’s path based on the given vehicle speed and steering angle. In the analytical vehicle model, two different moment arms were used for inboard and outboard wheels, and lateral and longitudinal load transfers were taken into account. In addition, the magic formula tire model was used to estimate the lateral force. The analytical vehicle model has been validated with a sophisticated ADAMS model, and it resulted in a good correlation with test data. Using the simple analytical model, a stochastic analysis was conducted to analyze the effect of the initial offset amount and the heading angle on the crash probability. Another stochastic analysis was also conducted to analyze the effect of a sensing error on the false negative rate (FNR) and the false positive rate (FPR). It was found that the initial offset amount and the heading angle played a key role in the crash probability, and only FPR was affected noticeably by a sensing error.Copyright