| dc.description.abstract | Recent years have seen considerable progress towards the goal of autonomous and unmanned ground vehicles which became essential for conducting military operations. These autonomous vehicles have the capability to operate and react to their environments without external control. Autonomous multi-wheeled combat vehicles are crucial for military applications which offer numerous leverages on modern battlefields. Applying autonomy features to such vehicles significantly increases its combat capabilities and expands its applications to work-day and night for risky missions compared with traditional manned ground vehicles. However, it is associated with some challenges because of their large dimension, heavy weight, and complex geometry. Therefore, the development of autonomous combat vehicles has become a cutting-edge research topic in robotics and automotive engineering.
This thesis focuses on the control issues related to applying autonomous features for the multi-wheeled combat vehicles due to their significant influence especially when navigating in the presence of obstacles. The primary concern of path planning is to compute collision-free paths. Another equally important issue is to compute a realizable path and, if possible, achieving an optimal path bringing the vehicle to the final position. For these purposes, the developed methodology considers the combination between the optimal control theory using Pontryagin's Minimum Principle (PMP) and Artificial Potential Filed (APF). In addition, a four-axle bicycle model of the actual multi-wheeled combat vehicle considering the vehicle body lateral and yaw dynamics is developed.
To generate the vehicle optimal path in real time, an Artificial Neural Network (ANN) model is proposed. The introduced ANN model allows the vehicle to carry out an autonomous navigation in real time with maintaining the path optimality by considering the vehicle parameters in terms of yaw rate, lateral velocity, heading angle and steering angle. Subsequently, a comparative study and performance analysis of the developed optimal path algorithm using PMP with Dynamic Programming (DP) method was carried out in order to guarantee the global optimum solution.
Determining the accurate vehicle position offers sufficient capabilities which increase the autonomy and safety features, especially in case of off-road locomotion. In this regard, a hybrid framework for positioning technique based on the integration of GPS/INS for combat vehicles is developed. The developed algorithm is able to provide an accurate and reliable vehicle positioning information, even if the number of visible satellites is less than four, due to the harsh vehicle operation environments.
In this work, a scaled multi-wheeled combat vehicle model was developed using system identification methodology. Different system identification methods are considered and applied to solve and identify this problem. An advanced control system in terms of fuzzy logic, robust, and PID control systems are designed. In addition, the Processor-In-the-Loop co-simulation (PIL) is considered, which permits and achieves a more realistic situation where the developed control algorithms running on a dedicated processor. The performance and effectiveness of the developed controllers are evaluated for vehicle heading angle tracking using different predefined heading angles. Furthermore, a comparative evaluation to assess the feasibility of the developed control algorithms is discussed. Finally, it should be stated that this work offers the first attempt in the open literature to control the scaled multi-wheeled combat vehicle using different advanced control techniques such as, fuzzy logic, 𝐻∞. | en |
