Max Robotics — Global Service Robot Directory + US Certification

About

Shashank Pathak’s research lies at the critical intersection of robotics, artificial intelligence, and formal verification, with a central focus on enabling robots to make safe, robust decisions under uncertainty. His most significant contribution is a unified framework for data association-aware belief space planning, which advances the state of the art by explicitly reasoning about data association—a source of uncertainty typically assumed to be perfect in existing planning approaches. This work, published in 2018 and cited 27 times, has become a foundational reference for robust active perception. Pathak is also deeply concerned with the safety of learned robot policies. His 2013 case study on the iCub humanoid (17 citations) demonstrated how to ensure demonstrably low collision probabilities when reaching objects near obstacles, bridging reinforcement learning with formal safety guarantees. He has further championed the use of probabilistic model checking as a tool for verifying robot control policies, arguing compellingly that formal methods are not merely a luxury but a requisite for safe adaptive robots operating in unstructured environments. Through this body of work, Pathak has established himself as a leading voice in the quest for robots that are not only intelligent but verifiably safe.