Systems Engineering Approach to Develop Guidance, Navigation and Control Algorithms for Unmanned Ground Vehicle

This research explores the development of a UGV capable of operating autonomously in a densely cluttered environment such as the tropical jungles or plantation estates commonly found in Asia.

Despite the growing popularity of unmanned systems being deployed in the military domain, limited research efforts have been dedicated to the progress of ground system developments. Dedicated efforts for unmanned ground vehicles (UGV) focused largely on operations in continental environments, places where vegetation is relatively sparse compared to a tropical jungle or plantation estate commonly found in Asia. This research explores methods for the development of a UGV that would be capable of operating autonomously in a densely cluttered environment such as that found in Asia.

Functional Decomposition to Perform Autonomous Navigation. See Technical Service Package below for high-resolution image.

The development of unmanned systems as a force multiplier in the armed forces has been gaining traction in recent years. This is largely attributed to the urgent need for reducing the operational risks faced by troops in a multitude of situations located within the various theaters of war. The pace for the development and deployment of unmanned ground vehicles (UGV) was, however, not keeping up with that of the other assets deployed in the aerial and naval realms of the military.

This research focuses on implementing a systems engineering approach for the rapid development of UGVs by exploring the requirements, available technologies, and the salient points on the development and tuning of the algorithm parameters that govern the guidance, navigation, control and target identification efficacy of the system.

Though platform specific, various tuning parameters were adjusted such as the linear velocity, angular velocity and look-ahead distance within the pure pursuit method, which was implemented for the UGV’s path following the algorithm. These parameters had an immediate effect on the UGV to accurately maneuver through a user-designated route. Any inaccurate tuning of the parameters resulted in the UGV exhibiting instability in its maneuver, often causing the system to veer off course or, even failing to reach its designated position.

As part of its autonomous capabilities, the UGV was designed to possess obstacle avoidance capabilities with the use of the vector field histogram algorithm. Parameters governing the detection and avoidance of obstacles were tuned in order for the system to perform adequately through the designated test route. These parameters included the detection limits of its sensor suite and the certainty thresholds of the algorithm to properly perceive the presence or discard the possibility of an obstacle within its path.

The implementation of the two algorithms – the pure pursuit method and the vector field histogram algorithm – allowed the system to maneuver efficiently through the user defined test route.

The last algorithm employed in this research was the use of feature recognition through image processing, which gave the UGV the capabilities to identify potential targets in its operational environment. Similarly, parameters requiring adjustment for the system to recognize the designated targets under varying light conditions and target profile exposure were also noted.

This research was based on a Pioneer 3AT robot platform equipped with light detecting and ranging (LIDAR) sensor, orientation sensor, image camera and an onboard mainframe computer as its primary hardware to detect, process and execute the autonomous behavior of the UGV. Software programs such as Ubuntu, MATLAB and Robotic Operating System (ROS) were also used in support of the system architecture to compute all variables for UGV operations.

Development and testing of the UGV prototype was conducted within a lab environment. This resulted in a verification set-up where the efficacy of the system was tested through varying light conditions and obstacle positioning to understand the performance and limitations associated with the respective methodologies applied in the system comprehensively. Pertinent behavioral changes associated with adjustments made to the tuning parameters were noted. This will allow for future iterations and development of the system to be carried out with relative ease using similar methodologies.

This work was done by Eng Soon Lim for the Naval Postgraduate School. NPS-0008

This Brief includes a Technical Support Package (TSP).
Document cover
Systems Engineering Approach to Develop Guidance, Navigation and Control Algorithms for Unmanned Ground Vehicle

(reference NPS-0008) is currently available for download from the TSP library.

Don't have an account? Sign up here.