Path-Following Control for Self-driving Forklifts based on Cascade Disturbance Rejection with Coordinates Reconstruction

The forklift is one of the most popular goods handling vehicles in logistic applications. Self-driving, as an emerging technology, is promising in simultaneously reducing the labor cost and improving the logistic efficiency for forklifts. In this paper, a cascade path-following controller for the self-driving forklift with rear wheel steering, based on a dynamic model, is proposed in the framework of active disturbance rejection control. To address the nonminimum phase behavior observed in forklift with rear wheel steering system, a coordinates reconstruction method is first applied. Then the lateral error, of the reconstructed coordinates to the target trajectory, is enforced to zero by manipulating the desired heading angle through the outer-loop controller. The heading angle is achieved in the inner-loop controller by adjusting the desired steering wheel angle. For both inner and outer loop controllers, all the effects of the discrepancy of the dynamic model from the real forklift are lumped into "total disturbances" for each loop, which are then estimated by the extended state observers. With the total disturbance rejected in the feedback loop, the plant is enforced to behave as a simple first order linear system which is straightforward to control. Experimental validation results show 20 cm maximum lateral error without need for controller gain scheduling, using low-cost positioning system with 15 cm positioning uncertainty. Frequency domain analysis is also provided for the evaluation of the robustness of the proposed controller.