Human intention-based collision avoidance for autonomous cars
- Resource Type
- Conference
- Authors
- Osipychev, Denis; Duy Tran; Weihua Sheng; Chowdhary, Girish
- Source
- 2017 American Control Conference (ACC) American Control Conference (ACC), 2017. :2974-2979 May, 2017
- Subject
- Aerospace
Bioengineering
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Engineering Profession
General Topics for Engineers
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Transportation
Automobiles
Trajectory
Autonomous vehicles
Roads
Predictive models
Prediction algorithms
- Language
- ISSN
- 2378-5861
This paper considers the problem of controlling an autonomous vehicle that must share the road with human-driven cars.We present proactive collision avoidance algorithms that take into account the expressed intent of human driven cars and minimize the need for sudden braking or other purely reactive sudden actions. The presented algorithm utilizes multi-stage Gaussian Processes (GPs) in order to learn the transition model for each vehicle given the intention of the vehicle's driver. It further updates the trajectory predictions on-line to provide an intention-based trajectory prediction and collision avoidance adapted to various driving manners and road/weather conditions. The effectiveness of this concept is demonstrated by a variety of simulations utilizing real human driving data in various scenarios including an intersection and a highway. The experiments are done in a specially developed driving simulation and a highly realistic third-party car simulator.