Investigating Control of Vision Based Autonomous Navigation in the Image Plane

This paper develops a novel integrated navigational system for autonomous vehicle motion control. Vehicle control is defined in terms of the required vehicle steering angle, rate of steering change and speed. This paper proposes predictive control in the image plane. The proposed predictive control enables the navigation on the desired path, reduces the control complexity and increases the application space for multiple types of vehicles. The paper investigates vehicle control stability; especially in scenarios containing varying curvature turns and variable vehicle speeds. The primary emphasis of this paper is on vehicle control rather than scene analysis. To demonstrate the proposed vehicle control, a computer vision based multi-lane detection algorithm is introduced. The control strategy is applied such that the vehicle maintains position within the lane boundaries. Stability of the control algorithm is tested and demonstrated in multiple scenarios using 3-D simulation results.

File Size: 3783K

Included in: V119-2

See papers presented at SAE 2010 Commercial Vehicle Engineering Congress, October 2010, Chicago, IL, USA, Session: Roadmap to Autonomous Vehicles (Part 1 or 2)

Purchase more technical papers and save! With TechSelect, you decide what SAE Technical Papers you need, when you need them, and how much you want to pay. Learn more >