Autonomous Navigation of an Unmanned Ground Vehicle in Cluttered Environment
Aaron Gaut · CSUN ScholarWorks (California State University, Northridge) · 2015
Unmanned Ground Vehicles (UGVs) extend operators' capabilities in many applications, especially in dangerous and hazardous situations. They are easy to deploy, fast to respond, and have a closer and more detailed observation of the environment. They are widely used for target detection, target recognition, and rescue missions. Applications include mapping, navigation, localization, ordnance disposal, and many other dull, dirty, and dangerous applications. The main advantage is that they pose no threat to human operators and are cost effective. However, in order to utilize their full potential, they are required to operate in areas that are full of obstacles such as in terrain, urban environments, battlefields, etc. Cal Poly Pomona is working on the development of fully autonomous UGVs that are capable of autonomous target acquisition and navigation to the target. Once the target is acquired, it is sent to the path planning algorithm. Vision system uses two stage optical object recognition using OpenCV2, and can detect as many targets as are visibly discernable. Primary target's latitude and longitude are calculated using current robot state. The route planning system is designed to safely direct the UGV to the target's location. The system incorporates obstacle and collision avoidance capabilities, and finds a navigable path to the target. The routing itself is done using a D*lite path planning algorithm that functions using a map that is populated using objects found by the LIDAR system. For initial testing, it is assumed that the area of operation is sparse, there is no instantly appearing targets, and that the area of operation is flat. Simulation and experimental results will be shown. In future, the research will be directed towards using this methodology for more complex scenarios where no assumption will be made on the area of operation and number of targets.