Demo: Blimp can recognize and move toward a green ball using color detection（the current blimp setup) Blimp is capable of seeking the ball by rotating itself Blimp is capable of moving toward an identified green ball Blimp is trying to capture the green ball using the capture basket Under...

The experiment I run in our lab this week is basically the altitude control test. The setup is as follows: Part of the data that I got from our system: The altitude info: The rotation in the altitude changes is also captured: Next I'll imporve the code to process the information(ro...

The video in the last post depicted a very slow detection of the ball. I also tested a CPU-trained model but, as expected, it was quite slow: To improve the performance of the algorithm, I replaced the original FasterRCNN model (torchvision.models.detection.fasterrcnn_resnet50_fpn()) with a mo...

I merged the ball detection code with the distance estimation code and added a camera component to make it live. The following video shows the code running live on CUDA: Model prediction is very slow on the CPU; I am currently checking out techniques such as quantization to optimize the model. A...

I'm still testing the property of different motor + propellers combination. Right now, since I also got the requirement from the AR blimp, I'll add one more test on to my previous experiments. The current experiment setup: So the properties of motor + propellers that I will cover in my exper...

I've had several flying tests this week to see how does feather board esp32 performs. Before I started the flying test, I need to configure the motor setup using inputs commands: The flying test of moving forward: an other angle of view: So this phenomenon(The forward motor completely...

I've finished the code that could by changing the input commands to get the correct FBUDLR(forward/backward/up/down/left/right) motor outputs: Use case: how to get the correct motor outputs after someone rewire the hardware The code is now in the github

This week, I've collected the green ball data with distance from two different scenarios: 1) Lab background 2) Conference room background The green ball data is constructed in the following tree structure: distance experiment |___background1 |___1280_1080 | 'time_Experime...

Over the last few days, I took some time formatting and labelling the newly obtained images of the ball at different distances from the camera. Following this, I passed the bounding box information through a regression model to detect the relation between the size of the bounding box (relative to th...

Since I'm starting to test different configurations this week, it would be easier if we could change the control accordingly. Right now, I've successfully adjusted forward/backward/up/down control even people have messed up with the wire connection in the wrong way. This is pretty useful since in...