Development and Demonstration of an On-board Mission Planner for Helicopters

. Mission management, including on-board replanning, is a task that can benefit significantly from automation. On-board replanning is required to respond to departures from nominal plan execution that result from imperfect knowledge of and temporal variability in the mission environment Automation is particularly valuable in the high-risk Nap-Of-Earth (NOE) environment, where crew warkloads for tasks of immediate concern (such as obstacle avoidance and h a t engagement) can be quite high. In this situation, an on-board, automated planning advisor can continuously monitor resource usage (i.e., fuel, ordinance, other expendables), assess risk along the cumnt mission plan and also suggest alternative plans that might better satisfy time, resource, and survivability constraints. The planning advisor requires an on-board environmental database of terrain, threat locations, and winds that is loaded preflight and updated during the mission by vehicle sensors and received communications. Also required is a mission database (also updated) indicating alternative objectives or bases, their locations, their relative values, and time of arrival constraints to ensure coordination with other force elements. Access to current vehicle navigation, fuel and expendables, and the absolute timt is essential to the planning advisor. Using a l l of this information in simple models to predict fuel and expendable use, time of d v a l and lethality risk, the planning advisor evaluates the current and alternative mission plans according to the probability of successfully accomplishing multiple mission objectives while satisfying mission constraints (exs. fuel, time, survivability, exclusion zones). Mission management tasks can be distributed within a planning hierarchy, where each level of the hierarchy addresses a scope of action, an associated time scale or "planning horizon", and requirements for plan generation response time. The current work is focused on the far-field planning subproblem, with a scope and planning horizon encompasing the entire mission and with a response time quired to be about two minutes. The far-field planning problem is posed as a constrained optimization problem and algorithms and structural organization are proposed for the solution. Algorithms are implemented in a development environment, and performance is assessed with respect to optimality and feasibility for the intended application and in comparison with alternative algorithms. This is done for the h e major components of far-field planning: goal planning, waypoint path planning, and timeline management. It appears feasible to meet Performance requirements on a 1.0 Mips flyable m e s s o r (dedicated to far-field planning) using a heuristically-guided simulated annealing technique for the goal planner, a modifred A* search technique for the waypoint path planner, and a speed scheduling technique developed for this project.