Max Robotics — Global Service Robot Directory + US Certification

About

Daniel Angelov is a robotics and machine learning researcher whose work sits at the intersection of robot learning, inverse reinforcement learning, and autonomous systems. His research addresses fundamental challenges in how robots can learn from human demonstrations, infer task objectives, and operate intelligently in dynamic environments. Among his most significant contributions is his development of probabilistic temporal ranking methods for learning reward functions from exploratory demonstrations — work applied to practical domains including robotic ultrasound scanning — demonstrating that robots can acquire complex search-and-discovery behaviors without explicit goal specification. His Vid2Param series of papers, garnering over 20 collective citations, introduced novel approaches to online system identification directly from video streams, enabling robots to reason about physical dynamics in real time. Further influential work on disentangled relational representations advances learning from demonstration by embedding structured inductive biases into policy and reward learning, making robot instruction more data-efficient and interpretable. Angelov has also explored causal analysis for extracting behavioral specifications from task demonstrations and hierarchical motion planning through change-point detection. With a growing citation record across multiple research threads, his contributions offer meaningful advances toward robots that can understand, learn, and generalize from human guidance in unstructured, real-world settings.