Last week, the team finalized InFrame’s design! Check it out under our Documents tab. Going over each aspect of our design really forced us to find holes and pinpoint weaknesses in our implementation strategy. By having to justify each design decision, we were able to drive our design further and are now ready to start building InFrame.
Some key design considerations that emerged/changed since finalizing our design review include:
- The use of a light ring would make for a much better user experience since we can give users feedback that they are being tracked, that the system lost them, where the system believes they are, whether or not it received a target from the phone, etc. Also, if something electronic doesn’t have LEDs, does it even do anything???
- Using some sort of honing mechanism to re-center the system to face forward on boot (since the pan servo is continous and doesn’t have any sense of absolute angles) like an IR or hall-effect sensor would be worthwile to consider adding.
- In the time between sending the current frame to the phone for the user to pick a target and actually receiving the target back from the phone, the bounding boxes sent will likely not be the same (i.e. potential targets move). As such, during that time, either multi-target tracking will have to happen (assigning IDs to the bounding boxes), or a user constraint of not moving while setting up the targets may need to be specified. The latter is not ideal but might be good enough for the InFrame MVP.
We’ve had great feedback on our design so far and are very excited to start building InFrame starting next week! Stay tuned!
0 Comments