Deep Predictive Models for Collision Risk Assessment in Autonomous\n Driving

2017/11/28 by Mark Strickland, Strickland, Mark, Georgios Fainekos +3 · 1 citation
Engineering · Psychology · #Autonomous Vehicle Technology and Safety #FOS: Computer and information sciences #Human-Automation Interaction and Safety #Robotics (cs.RO) #Traffic and Road Safety

paper · pdf · doi:10.48550/arxiv.1711.10453

openalex publication_date 2017/11/28 · openalex created_date 2022/10/07 · openalex updated_date 2026/07/28

Abstract

In this paper, we investigate a predictive approach for collision risk\nassessment in autonomous and assisted driving. A deep predictive model is\ntrained to anticipate imminent accidents from traditional video streams. In\nparticular, the model learns to identify cues in RGB images that are predictive\nof hazardous upcoming situations. In contrast to previous work, our approach\nincorporates (a) temporal information during decision making, (b) multi-modal\ninformation about the environment, as well as the proprioceptive state and\nsteering actions of the controlled vehicle, and (c) information about the\nuncertainty inherent to the task. To this end, we discuss Deep Predictive\nModels and present an implementation using a Bayesian Convolutional LSTM.\nExperiments in a simple simulation environment show that the approach can learn\nto predict impending accidents with reasonable accuracy, especially when\nmultiple cameras are used as input sources.\n

Cited by

• How Safe Will I Be Given What I Saw? Calibrated Prediction of Safety Chances for Image-Controlled Autonomy

Related

• Real-time Prediction of Intermediate-Horizon Automotive Collision Risk
• Intention-based and Risk-Aware Trajectory Prediction for Autonomous Driving in Complex Traffic Scenarios
• Risk-aware controller for autonomous vehicles using model-based collision prediction and reinforcement learning
• Multimodal Trajectory Predictions for Autonomous Driving using Deep Convolutional Networks
• Vehicle Trajectory Prediction based Predictive Collision Risk Assessment for Autonomous Driving in Highway Scenarios