Navigating in cluttered or crowded spaces demands a paradigm that seamlessly combines planning, feedback control, and real-time sensing. Drawing inspiration from natural systems, the integrated planning and control (IPC) strategy merges trajectory design and execution into a single cohesive framework, thereby reducing the latency and rigidity common in conventional decoupled approaches.
In this work, building upon the original IPC strategy by Veejay Karthik at IIT-B , we extend invariant-set-based navigation to 3D underwater environments. By employing motion sets derived from non-linear feedback control laws, the system maintains naturally induced safe trajectories, swiftly responding to environment changes. Parameterized motion sets, along with continuous onboard sensing, enable the AUV to adapt to uncertain surroundings in real time. A manuscript detailing these advancements is currently under preparation for submission to a top robotics journal.
Below are two demonstration videos of an AUV navigating a cluttered 3D environment:
These scenarios highlight how a tightly coupled planning and control approach leverages continuously updated feedback to maintain robust collision avoidance, even as new hazards arise in real time. My contribution specifically focuses on adapting and validating these techniques in the unique domain of undersea robotics, where reduced visibility, unpredictable fluid dynamics, and multidimensional motion further amplify the need for integrated solutions.