Date of Award
Master of Science in Electrical Engineering (MSEE)
Electrical, Computer, and Biomedical Engineering
With the recent development of machine learning and neural networks, different applications have been developed to improve the intelligence of unmanned robotic systems. In this thesis, we present a novel method for controlling an unmanned ground vehicle (UGV) by using a new machine learning technique, called deterministic learning , to learn and recognize four specifically designed body language, which represent four corresponding moving directions (i.e., left, right, up, and down) of the controlled UGV. The Microsoft Kinect sensor is employed to collect the human body skeleton data, including (x, y, z) coordinates of the human arm joints, from which four specifically-designed features are extracted for neural network training. The discrete-time deterministic learning algorithm is then utilized to train the radial basis function neural networks (RBFNNs). The dynamics of the human arm waving motion is guaranteed to be accurately identified, represented, and stored as an RBF NN model with converged constant NN weights, which facilitates the rapid recognition purpose in the testing phase. In the testing phase, a set of estimators are built based on the database established in the learning phase, so as to conduct real-time rapid recognition of new in-coming gesture commands. The smallest error principle is used to decode the human intention, the decoded result will then be sent to the UGV through TCP/IP to control its moving directions. A full-integrated graphical user interface (GUI) has been developed based on Python programming language to demonstrate the effectiveness of the proposed approach and illustrate the proposed experimental results.
Chen, Xiaotian, "UGV DIRECTION CONTROL BY HUMAN ARM GESTURE RECOGNITION VIA DETERMINISTIC LEARNING" (2019). Open Access Master's Theses. Paper 1487.