Adaptive preview tracking fuzzy control algorithm for tracked vehicles

Tracked vehicles have demonstrated superior passability in hilly and mountainous terrains. Among them, the control system of path tracking has been focused mainly on wheeled agricultural equipment rather than tracked vehicles, particularly with single-sided braking tracks. In this study, the fuzzy control of adaptive preview tracking was proposed to improve the parallel control and tracking accuracy with the lower turning count of tracked vehicles. The 3B55 tracked transport vehicle was modified as the experimental platform from Jiangsu Zhushui Agricultural Machinery. A preview tracking model was constructed via the kinematic analysis of single-sided brake-tracked vehicles. Efficient parallel control was achieved in the steering and straight-line travel within the same control cycle. A multiple input/output fuzzy controller was designed with the lateral and heading deviation as the inputs, while the motion distance and turning ratio as the outputs. Furthermore, an adaptive line-of-sight solution was introduced with the improved sparrow search algorithm (SSA) to enhance the accuracy of path tracking with less turning count. The chaotic factor and clustering scavengers were also added to the sparrow population, in order to improve the convergence accuracy and susceptibility to local optima. This adaptive front view distance(L_os) solution with the improved SSA was analytically determined as the optimal L_os, considering the lateral deviation and steering path angle constraints of the vehicle's current state. Simulation and field experiments were carried out to evaluate the tracking accuracy and turning count. Simulation results demonstrate the convergence speed was significantly improved during the early iterations of the improved Sparrow Search. Moreover, the superiority of the improved SSA was validated to successfully escape the local optima with the optimal fitness value after 37 iterations. The on-line modelling test was implemented with the L_os of 0.8, 1.5, 2, and 3 m. The lateral deviation increased in the same control cycle when the vehicle was reduced the heading deviation. A longer time was required for the vehicle to complete the on-line, with the increase of the L_os. The accuracy of vehicle tracking decreased gradually, but the number of turning counts was also reduced. An optimal L_os was obtained in the fuzzy control of adaptive preview tracking in real time, according to the current position and steering path angle of the tracked vehicle during path tracking. When tracking multi-angle planned paths, the turning count was 89 times, with an error area of 1.74 m². The field experiments were conducted at the Panax Notoginseng Planting Test Field of the Yunnan University Traditional Chinese Medicinal Materials Mechanization Engineering Research Center. The test field was measured as 40 m in length and 6 m in width. The accuracy of path tracking decreased for the tracked transport vehicle in the field, due to the unevenness of the terrain. However, the trends of tracking accuracy and turning count with the L_os were consistent with the simulation. Both the turning count and path tracking accuracy decreased gradually, as the driving speed increased, especially in the tracking fuzzy control with fixed L_os. When the vehicle tracked at the speeds of 0.14, 0.47, and 0.83 m/s, respectively, the fuzzy control of adaptive preview tracking reduced turning count by 13.59%, 9.87%, and 11.25%, respectively, and the error areas by 19.93%, 48.48%, and 54.59%, respectively. The field and simulation proved that the adaptive predictive tracking fuzzy control for tracked vehicles can be expected to improve the tracking accuracy of the vehicle and the number of steering control times in the field tracking. The tracking trajectory was closer to the planned path, indicating better performance. This study can provide innovative ideas and technical support to the automatic navigation of agricultural machinery in the single-sided brake-tracked vehicle.