Path planning and real-time control optimization method for unmanned ground robots in complex terrain

Unmanned ground robots (UGRs) often work in rough and irregular terrains where traditional path planning and control methods are not reliable. This study develops a combined method using an improved genetic algorithm (IGA) and an enhanced sliding-mode controller (ESMC). The IGA creates smoother and shorter paths by adding terrain slope, curvature, and obstacle risk into the cost function. The ESMC keeps the robot stable under external disturbances with the help of a disturbance observer. Experiments were carried out in a 10 m × 10 m outdoor test field that included uneven surfaces and random obstacles. Compared with the Dijkstra method, the combined system shortened travel time by 18.7%, limited path deviation to ±4.2 cm, and kept attitude error within 1.5°. These results show that linking path planning with real-time control improves tracking accuracy and stability on rough ground. The method can be used for rescue, mapping, and inspection tasks, and future work will test its performance with moving obstacles and soft soil.