Automated Co-Design of Vehicles and their Teaming Operations for Optimal Off-Road Performance

Project begins 2023.

This work proposes to create methods to co-design vehicles and their teaming operations for optimal off- road performance in combat. The proposed design problem defines a family of optimization problems to find the minimal resources needed to implement a given functionality to each vehicle while maintaining optimal performance and overall team success, with a focus on combat operations. Aspects that are subject to co-design include fleet size, vehicle-specific physical attributes, and vehicle-specific operation algorithms (path planning and communication) for the autonomous vehicles. Thus, this research aims to answer the following fundamental research questions:

RQ1: How to automatically generate parametric vehicle designs within a predetermined parameter space, identify effective tactics for each parametrically generated vehicle in a representative sample of expected combat terrains, and compare effectiveness in simulated combat against each other? The novelty is in the creation of specialized (deep learning) algorithms using generative networks for vehicle design that also identify the optimal operation of vehicles.

RQ2: How to create human-autonomy teaming algorithms that address an intelligent demand (adversary) for automatically generated vehicle designs? The novelty is in a unique extension of algorithms for communication and dynamic teaming building on outcomes of past ARC research.

RQ3: How to structure a vehicle design problem in a modular way, in which each different capability option can be adaptively “plugged in” in a modular and dynamic environment automatically generated? The novelty is in the creation of dynamic teaming algorithms integrated with modular vehicle design algorithms building on outcomes and deliverables of past ARC projects.

Publications from Prior Work closely related to the proposed project:

1. Sapunkov, Oleg, “Historical Trends and Parameter Relationships in the Design of Armored Fighting Vehicles,” GVSETS 2020.
2. Wu, H., Ghadami, A., Bayrak, A. E., Smereka, J. M., & Epureanu, B. I. (2021). “Impact of Heterogeneity and Risk Aversion on Task Allocation in Multi-Agent Teams.” IEEE Robotics and Automation Letters, 6(4), 7065-7072.
3. Wu, H., Ghadami, A., Bayrak, A. E., Smereka, J. M., & Epureanu, B. I. (2022, May). “Task Allocation with Load Management in Multi-Agent Teams.” In 2022 International Conference on Robotics and Automation (ICRA), IEEE, 8823-8830.

#5.23