In-Flight Adaptive PID Sliding Mode Position and Attitude Controller

Future spacecraft operations, such as debris capture and autonomous in-orbit servicing, will require spacecraft to interact with objects with poorly defined inertial properties which introduce a large degree of uncertainty into the dynamics of such operations. To handle this uncertainty, this paper describes the development and validation of a mass property-resilient controller for position and attitude control of a free-flying spacecraft. Specifically, the proportional-integral-derivative (PID) sliding mode controller (SMC) was developed to account for inaccurate knowledge of mass properties of small spacecraft, using sliding mode variables for each axis and an adaptive determination of the appropriate integral and derivative gains to achieve a commanded motion. The controller was validated both in simulation and in ground-based tests on the SPHERES (Synchronized Position Hold Engage Reorient Experimental Satellites) platform, small free-floating autonomous satellites used to study precision navigation and maneuvering. The SPHERES was commanded to follow a specified series of maneuvers using the PID SMC, to assess the convergence time of the sliding variables, gain values and commanded state parameters in each maneuver. The PID SMC performed well in simulation, with the sliding variables associated with position converging to the desired zero value within the span of one maneuver, while the attitude controller exhibited some chattering. In hardware tests, the position sliding variables exhibited slight chatter, consistent with slow convergence to the desired zero value, whereas the attitude sliding variables had significant chattering. The same tests were run using a static PID controller designed with inaccurate mass property knowledge, and the results were compared to those of the PID SMC. From this comparison, it is evident that the PID SMC offers superior performance under mass property uncertainy. Future work will attempt to address the PID SMC's chattering issues with the introduction of a boundary layer in the attitude sliding variable.