Open Access Thesis
Date of Award
Electrical and Computer Engineering
Controlling a ground vehicle using only onboard and aerial camera views, can be a tedious task, as it is challenging for the operator to estimate information about the robots location and orientation using only these view. However the visual data can be sufficient for computerized calculation of this information even in the absence of other sensors. This approach is of interest to the military because the use of passive sensors instead of active sensors (e.g. LIDAR) would be easier and more reliable and would not give off detectable signatures.
The goal of this research was to control a ground vehicle using onboard and aerial camera views. No other sensors like an Infrared camera or IMU were used to track the robot. In addition an augmented reality-based tracking capability was developed for the ground robot using an aerial vehicle. Analysis and efficiency testing was conducted on the implemented tracking and control algorithms demonstrating that the ground vehicle can be tracked by the aerial robot using augmented reality. Moreover the augmented reality control system enabled the ground vehicle to be accurately directed to different locations using the onboard and aerial camera views. Although the research used a specific robotic platform the Surveyor SRV1, the concept can be implemented on any other ground vehicle or potentially be applied to a swarm of robots.
Lowalekar, Vishal, "Control of a single ground vehicle using aerial and onboard camera views" (2011). Wayne State University Theses. Paper 127.