Performance evaluation of sensor combinations for mobile platoon control.

摘要

Autonomous relative navigation of vehicles may soon be feasible. The concept of Collaborative Driving Systems (CDS) involves linking several vehicles together in a platoon. This will have many benefits, including increasing road capacity, improving safety, and reducing driver fatigue and stress. This thesis is focused on examining various positioning sensors for potential use in a CDS. Firstly, the use of GPS as a relative positioning sensor is examined. Tests were conducted using four instrumented vehicles, each equipped with a precise GPS position and heading determination system. This enabled the relative position and velocity estimation between vehicles as well as between antennas with a constant inter-antenna baseline, using carrier phase and Doppler observations. Relative position accuracy was shown to be within a few centimetres, while relative velocity was accurate to a few centimetres per second. Secondly, various sensors were mounted on mobile robots. The lead robot was manually controlled, while another robot was left to autonomously follow. The sensors on the robots include GPS, a digital camera, and a laser scanner. Results show that GPS gives very high accuracy distance measurements but can not consistently or accurately provide angle. The digital camera provides continuous distance and angle measurements, though with less distance accuracy than GPS. The laser scanner provides distance and angle measurements of high accuracy, except often has difficulty identifying the target. Combining GPS with either auxiliary sensor resulted in superior performance, by reducing time to fixing ambiguities, and improving accuracy during GPS data outage and float ambiguity positioning.