Improved Lateral Stability for Unmanned Ground Vehicle Convoys

Unmanned convoying, where a manned or unmanned leader is followed by multiple unmanned vehicles, is a mode of operation that maximizes available manpower for transporting goods and materials. In many situations, it is desirable to have as many vehicles in the convoy as possible to increase efficiency. Currently, problems in the lateral stability of the convoy can occur as more and more vehicles are inserted into the convoy. ASI completed a Phase II project for lateral improvements to the convoy when many vehicles are used. The work done has shown that when the vehicles communicate certain information, the overall performance of the convoy can improve. The purpose of this proposal is to build on the current success of the Phase II project and extend the algorithms to 1) GPS-denied operation 2) obstruction management, 3) improved convoy communications, and 3) full convoy testing using 1 manned leader and 5 autonomous follower vehicles. This project will be completed by Autonomous Solutions, Inc. (ASI) and Auburn University. We propose to investigate several approaches to accomplish GPS-denied convoying with multiple unmanned vehicles following a manned leader. These approaches include a non-line-of-sight following mode that depends only on camera and INS approach, a line-of-sight mode where UWB and INS is used for both lateral and longitudinal control, and a line-of-sight mode where longitudinal control is accomplished with UWB ranging and lateral control is accomplished with an enhanced road following technique. Previous work done at ASI and Auburn that allows vehicles to track and avoid obstacles will be continued and improved and applied to the GPS-denied convoy scenario. We will develop efficient and reliable vehicular network protocols that support the forwarding of critical control information among trucks in long vehicle convoys. The vehicle network protocols will automatically reconfigure to adapt to real-time changes in the vehicle topology, dropout and joining of vehicles from and to the long convoy, large number of vehicles and arbitrary sizes of the vehicles. This network protocol will support transmission of critical control data between multiple vehicles in long vehicle convoys that requires multi-hop forwarding of control data over a vehicular ad hoc network (VANET) from the lead vehicle to all the rear vehicles. The algorithm development and test scenarios are done with representative environments and vehicles. Each of the algorithms shall be tested in simulation and on autonomous vehicles.