An Augmentation of G-Guidance Algorithms
- Saturday, 01 January 2011
This augmented algorithm can be used in small-body proximity operations utilizing model predictive control with a need for safety from surface-constraint uncertainty.
The original G-Guidance algorithm provided an autonomous guidance and control policy for small-body proximity operations that took into account uncertainty and dynamics disturbances. However, there was a lack of robustness in regards to object proximity while in autonomous mode. The modified G-Guidance algorithm was augmented with a second operational mode that allows switching into a safety hover mode. This will cause a spacecraft to hover in place until a mission-planning algorithm can compute a safe new trajectory. No state or control constraints are violated. When a new, feasible state trajectory is calculated, the spacecraft will return to standard mode and maneuver toward the target. The main goal of this augmentation is to protect the spacecraft in the event that a landing surface or obstacle is closer or further than anticipated. The algorithm can be used for the mitigation of any unexpected trajectory or state changes that occur during standard mode operations.
In order to have the G-Guidance algorithm detect an unsafe condition, it required some modification. This modification provides a policy to safely maneuver the spacecraft between its current state and a desired target state while ensuring satisfaction of thruster and trajectory constraints, along with safety constraints. In standard mode, this modification brings the spacecraft from its current position closer to its target state. In safety mode, the algorithm maintains the spacecraft’s current state at zero velocity. Since the safety mode is designed to be temporary, the destination location in this mode is also temporary, and once a new destination location is provided, the spacecraft returns to standard mode.
The G-Guidance algorithm uses both a planned trajectory (feedforward) and a control policy (feedback), along with sensors to monitor actual spacecraft state. The feedback is designed to ensure that the spacecraft stays within a specified proximity to the feedforward. The feedforward is designed to achieve the goals of each mode: hover for safety mode and maneuver toward target for standard mode. By giving the spacecraft the ability to re-compute its trajectory on-the-fly in response to local conditions, minimization of fuel usage is provided. The original G-Guidance algorithm provides robustness to uncertainty affecting the dynamics. The safety augmentation provides a form of state-constraint robustness, which further mitigates risk.