As in previous years, I competed in the Japan Innovation Challenge this year as part of the TAP-J team. We took home 3.3 million yen in prize money (about $22K USD) which was a good result. This was much less than last year but mostly because other teams (often also using ArduPilot) did better so the prize money was divided up more which is all good of course.
As in previous years there are three stages to the event:
- Stage1: Search. The organisers hide two heated mannequins in a 1km^2 area of the total competition area. Teams are given 80min to find them starting from around 6pm. Vehicle are setup at 2pm and are operated from a remote location and cannot be touched until after all teams have completed the stage
- Stage2: Delivery. The organisers provide the lat, lon location of a mannequin and the teams are given 30min to deliver a 3kg rescue package to between 3m and 8m (too close is also a fail!). This takes place between 10pm and 12am but again the vehicles must be setup by 2pm
- Stage3: Rescue. The organisers provide the lat,lon location of a 50kg mannequin and teams have 5 hours to retrieve it and bring it back to the start area. The mannequin must be handled gently so that onboard IMUs do not receive more than 3G of shock. This takes place between 11am and 4pm
.. but this year the competition was shortened to 2 days (previously it was 3) and the “remote area” (from where the operators control the vehicles) was in Tokyo for stages 1 & 2 which is about 900km from the main event site in Kamishihoro, Hokkaido, Japan. This meant we needed to divide our team up so intra-team communication more difficult but from a technical point of view, it didn’t make a huge difference because we’ve been using LTE/4G telemetry for years in any case.
Here’s what new technology the team used this year:
- BlueOS replaced our custom companion computer system from earlier years giving us:
a) live streaming video from our search and delivery copters which made it even easier to find the mannequin quickly. Below is a screenshot from the 2nd day’s delivery attempt. We could see the mannequin (shown in red) and also see the payload (see green arrow) giving us confidence and potentially allowing us to manually adjust where the payload was placed.
b) NTRIP RTK paired with Here4 GPSs gave us absolutely fantastic position accuracy. In previous years we suffered from the vehicle sometimes landing off the landing pad (see 2024 video here) but this year the vehicle landed exactly back on the pad within cm of where it took off from. Below is a pic and you can see how close the takeoff footprint is to the vehicle’s landing location
c) easy download of pictures using this BlueOS Camera Download extension. This meant we could download the images easily during flight and start the analysis early so we often already knew where the mannequins were even before the vehicle completed it’s mission - We replaced our search copter with a larger vehicle identical to our delivery vehicle. It was frankly much larger than necessary but it has >1hr flight time and having a single platform with identical tuning, etc made maintenance easier.
- We replaced our custom rover frame with a modified Komatsu PC12UU-2 mini excavator. We used linear actuators paired with short-range lidar that could return the absolute length of the actuator thus giving us absolute position control rather than only relative. Sorry I don’t have more infomation on this setup (I took responsibility for the multicopters, not the rover)
I’d like to thank all these people that contributed to this!
- Kamishihoro local government for putting on such a great event!
- TAP-J team members including RtoS and QuKai for all their efforts
- AttracLab for providing the multicopter frames
- Siyi for providing the ZT6s which we used on both multicopters and provides absolute temperatures both in real-time and in the final images. This is a critical feature that every camera used for S&R should have
Here’s a video recorded from the 2nd day’s delivery
Finally congratulations to Yutaka Kaizu from Tokyo Univesity whose team brought back the mannequin safely and claimed the big 20 million yen ($130K USD) prize!
I’ve known Kaizu-sensei for years and they also use ArduPilot actually
. This means there may not be a Japan Innovation Challenge in 2026 but then again, there’s always HEDC.