Research Projects

Introduction: The proposal stems from the necessity of automation in military surveillance, crop monitoring of large farmlands, assessment of inaccessible disaster zones (e.g. wild fires), and pollution assessment of large water bodies (eg. oil spills). The research proposal aims to efficiently compute optimal strategies for surveillance and area coverage by a team of fixed-wing unmanned aerial vehicles. These vehicles have the advantage of fuel efficiency and endurance over rotary-wing aerial vehicles. But the planning for these vehicles is particularly challenging due to constraints such as constant speed for maintaining a constant altitude, non-holonomic slipping constraints, and bounded angular velocity input. The current literature on surveillance and coverage assumes very simplified vehicle models that violate these constraints and do not guarantee optimality. This limits the practical implementation of such strategies on real vehicles. In this proposal, we aim to develop and implement optimal surveillance and area coverage strategies for fixed-wing UAVs modeled as Dubins vehicles while simultaneously minimizing the following two performance indices: 1. Total time for area coverage, thereby achieving faster coverage. 2. Minimizing fuel consumption of each vehicle, thereby improving the endurance of these individual UAVs. Approach: We will primarily use optimal control theory and the reachability set of Dubins vehicles, combined with tools such as control barrier functions to handle constraints effectively. Cooperative behavior among UAVs can be enforced using either coupled cost functions or graph-theoretic models to represent information sharing among connected UAVs. While optimal control theory or optimization-based approaches offer a robust framework for addressing constrained motion planning, their applicability in real-time scenarios is often limited by the curse of dimensionality and the general difficulty in solving the Hamilton-Jacobi-Bellman equation analytically. To address this, one possible direction is to first characterize the globally optimal solution for a small number of vehicles using only the first-order necessary conditions for optimality. This characterization is expected to reduce the search space for the optimal strategy while providing geometric and analytical insights into the problem. These insights can inform better heuristics and serve as building blocks for tackling more complex coverage problems. Experimental Developments: We will establish a test-bed to validate the proposed strategies on unmanned aerial vehicles in laboratory. The insights from the experiments will be further used to deploy these UAVs in real world scenario. Once deployed, the developed system will lead to efficient and rapid surveillance and area coverage by fixed-wing UAVs that can be deployed for the aforementioned civilian and military domains.