Jonathan Cruz
Papers
5
Total Citations
60
H-Index
3
About
Jonathan Cruz is a leading researcher at the intersection of robotics, embedded systems, and artificial intelligence, with a focus on enabling autonomous capabilities on severely resource-constrained platforms. His major contributions center on bringing deep reinforcement learning (deep-RL) to nano drones—tiny, low-power quadcopters—for real-world tasks like autonomous source seeking. His most cited work, "Tiny Robot Learning (tinyRL) for Source Seeking on a Nano Quadcopter" (2021, 26 citations), demonstrates how application-specific system and observation feature design allow a deep-RL policy to run fully onboard a nano quadcopter’s microcontroller, achieving high-performance autonomous navigation. This builds on his earlier 2019 paper (20 citations), which pioneered the same concept. Cruz also addresses the computational bottlenecks of 3D perception at the edge with his work on the OMU accelerator (2022, 10 citations), a probabilistic 3D occupancy mapping engine that enables real-time OctoMap on resource-limited robots. His research has earned over 60 citations, and he is recognized for pushing the boundaries of what tiny, energy-constrained robots can accomplish autonomously, with direct implications for search-and-rescue, environmental monitoring, and swarm robotics.
Research Focus
Key Achievements
Top Papers
- 1Tiny Robot Learning (tinyRL) for Source Seeking on a Nano Quadcopter26 citations · 2021
- 2
- 3
- 4RoboRun: A Robot Runtime to Exploit Spatial Heterogeneity2 citations · 2021
- 5