Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 50 | Autobrake on Bicycle |
Gary Gong Taiqing Ling Weichen Qi |
Xihang Wu | design_document1.pdf final_paper1.pdf proposal1.pdf |
|
| ## Problem Statement: Self-driving is always a topic heatedly discussed by people and more and more AI products have been invented these years, like lane keeping, blind spot detection, etc. Most of these products are designed for cars. However, we think it is equally important to have some AI products on bicycle in order to guarantee people’s safety. Sometimes, accidents happen because cyclers are distracted and they do not notice a stop sign. Therefore, to prevent such cases, in this project we want to design a STOP sign detector that can brake the bicycle automatically. ## Overall solution: Here is how our project will work. It will first detect a STOP sign, and generate some waning sounds to remind people. Then, if people do not take action after receiving the warning, the auto brake will be triggered to prevent accident. The project will be split into the ‘cycle’ part and the ‘helmet’ part. On the ‘cycle’ part, it will be able to detect a stop sign when it is closed enough. We plan to use a camera module and use the ‘template match’ algorithm to detect a stop sign. We can also determine the distance of detection by adjusting the size of convolution kernels. It does not need to accurately measure how many meters the stop sign is away from the bicycle. It only needs to determine a threshold distance that for triggering autobrake. We will do several tests to adjust the kernel size so that the threshold is set reasonably. Once a STOP sign is detected, it will send a signal to the helmet, and warn people about it. We will also measure the speed of the bicycle so that if the speed is higher than some value while the STOP sign is closed enough, which means the driver did not take action after hearing the reminder, the auto brake will be triggered. ## Components list: # Cycle: - arduino board - zigBEE: wirelessly connected with the helmet - camera: capture image in front of the vehicle - hall effect sensor: installed on the wheel to detect speed - servo with high torque: trigger the bicycle brake ## Helmet: - Lilypad: wearable micro controller - zigBEE: paired with the one on bicycle - speaker: warning sounds ## Criterion of success: We consider our project a success if: 1. It can generate warining sound on the helmet glass after STOP sign is detected. 2. The autobrake is triggered if the wheel is not slowing down within specific distance. |
|||||