Keynote: Python vs Orangutan
summary
A drone system detects implanted radio pings to locate six orangutans across ~2000 km2 by flying lawn‑mower patterns, collecting samples with a software-defined radio and a single‑board computer, and running on‑board processing in Python with numpy. Implants ping every two seconds and handheld range is ~200 m; drone tests reach 120 m altitude, ~20–30 m localization accuracy, ~20 min endurance. Prototype proves the concept in Borneo but endures heat, crashes and a lost airframe; more sensitivity and terrain following are needed.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:48Orangutan tracking problem
- 8:29Drone tracking proposal
- 12:22Python drone system
- 15:08Signal processing logic
- 17:57UK proof of concept
- 21:52Borneo field testing
- 26:08First blind tracking success
- 30:28Challenges and drone loss
- 31:33Lessons learned
- 34:21Conservation hope
related talks
33mPython vs Orangutan
15mThe Internet of Wild Things: Acoustics, AI and Conservation
43mTo kill a working drone: flying a drone with twitter, bananas & Web APIs
45mFlying a drone with gestures, bananas & Web APIs
33mFlying a drone with twitter, bananas & Web APIs
10mDeep Learning in the Wild: Embedded PyTorch for...
26mMaking an Impact with Duct Tape
17mTechnology: A Changemaker in Environmental Protection
