SkyAgent improves drone search with smarter target memory
SkyAgent: Revisable Target Memory for Continuous Aerial Search
Robotics
Summary
Searching for things from a drone’s camera can be tricky because targets look different from different angles, and sometimes a target thought to be found is actually not there. The authors created SkyAgent, which remembers candidate targets but updates or removes them when new information contradicts earlier guesses. Their system links every movement decision to the specific target idea it came from, so drones don’t waste time chasing false leads or miss temporarily hidden targets. Tests on real and simulated drone data show this method finds targets more reliably and reduces unnecessary travel.
What this means in practice
- •For drone operators: Improve continuous aerial search missions by reducing false target pursuits and increasing success with dynamic target memory revision.
- •For security and surveillance teams: Enhance UAV patrol reliability by minimizing unnecessary travel caused by outdated target hypotheses and managing temporary occlusions.
Authors
Kaibo Yang, Xingyu Wu, Enze Zhu, Yajie Chai, Xuyang Li, Jiahui Cui, Xiaoxuan Liu, Lei Wang, Zhan Chen
Abstract
Continuous aerial search builds target hypotheses from sparse, viewpoint-dependent observations. When later evidence contradicts a stored hypothesis, however, the candidate can remain available to the planner and continue triggering approach or verification actions. We call this trajectory-level failure false candidate persistence (FCP). We equip SkyAgent with Revisable Target Memory (RTM), which associates evidence with persistent candidates, discounts correlated views, and revises each candidate through Tentative, Confirmed, and Rejected states. The key interface binds every generated waypoint to its source candidate: rejection therefore invalidates stale goals, while conservative reactivation protects temporarily occluded targets. Across four splits from the CityAVOS, UAV-ON, and OpenUAV benchmarks, our SkyAgent equipped with RTM outperforms all competitors, improving SR by 2.90--6.31 points while substantially reducing Behavior-FCP and post-contradiction travel with a low false-rejection rate. Across 35 matched real-world DJI scenarios per method, RTM consistently improves task success and behavioral reliability, further validating its state-to-action correction under physical deployment conditions.