A Year In A Rainforest: Engineering For Survival
summary
Remote ocean and rainforest devices run months by using an FPGA for real‑time filtering and a microcontroller that stores only metadata to SD cards, backed by multiple selectable battery stacks, capacitors for power interruptions and coulomb counting for load testing. Error handling records three errors in eeprom, appends startup and shutdown report files (plain English), and uses software + hardware watchdogs. Bootloader stays minimal (flag + external flash), firmware‑over‑the‑air tests against oldest/newest versions, and the project logs 172 releases in four years.
AI-written summary · about
chapters (28)
- 0:00Intro
- 3:30Deployment Challenges
- 5:48Dolphin Bat Detectors
- 8:08The Explorer
- 17:30Planning
- 17:54Diary
- 18:35Logs
- 19:37Sustenance
- 19:57Packing food
- 20:29Power
- 23:03Low Power
- 25:28Handling Problems
- 26:40Show Of Hands
- 28:48Live Code
- 30:30Live Code 2
- 32:22assert
- 34:08undefined behavior
- 35:38standard strings
- 36:28interview question
- 37:05option 3 problems
- 38:28option 4 problems
- 41:43watchdogs
- 43:25software watchdogs
- 45:15implementing changes
- 48:41error reporting
- 51:23subsystem
- 52:54report files
- 57:48infield updates
related talks
41mEight Years of Farming; Is EveryboArDy Happy?
48mRunning an Undersea, Robotic Laboratory on a Fixed Energy Budget
01h 00mIntermittent Computer Systems on Earth and in Space
06mBFree: Enabling Battery-free Sensor Prototyping with Python
24mPython Untethered: Building Robust Embedded Systems
18mTinySDR: Low Power SDR Platform for Over the Air Programmable IoT Testbeds
23mBuilding An Over Engineered Weather Station
35mEmbedded Linux in EOD Robots - Lessons Learned