Conclusion and future developments

To achieve a fully functional truck driver wake-up device was challenging; it required developing a complete set of python codes. Image processing method was used for detecting real-life traffic light signals. Obtaining the correct tools and functions for the detection was a key to success during this project, and it was achieved. The OpenCV functions were used to meet the goals, it required a lot of installation on the Raspberry Pi, which took around 6 hours. However, using the OpenCV functions made it very easy for the detection.

Traffic light detection is complicated task due to the visual noises around. In order to overcome this problem, a method was proposed for detecting only green circles. The proposed method using Raspberry Pi camera module was fully modular and capable to recognise the green signal at most of the adverse conditions. Testing was performed to verify the effect of the proposed method.

The Android web apps were developed successfully, live streaming of the camera was easily accessed, and the Pushbullet app was working completely fine during the whole testing procedure. Using a power bank as a power source made it easier for the user to find the battery life left through the display screen. A user manual is available, it shows that the device can be used by anyone by just following 7 steps. That meets one of the requirements that were set for the device; “Easy to be used/trained on, manual will be provided”.

To conclude, reflecting on the goals and requirements set for the wake-up device, majority of the requirements were met. The project is a success because it proves that the principles employed for the project work well. However, many things can be improved for future developments. As mentioned above in the report, the Raspberry Pi camera module wire extension was needed to separate the camera from the Raspberry Pi and the power bank due to the weight. As a future improvement, a USB web camera will be used instead of the Raspberry Pi camera module, as it does not require any wire between the device and the camera. Also, viewing live stream of the camera and the notification are not there on the same app or page, so for future work both notification and live streaming will be available through the same app. Adding a solar panel to generate the flow of electricity to the device will be considered as well. Through the report, the device was tested on extreme weather conditions except for snow and fog (the weather conditions in the UK are not supportive). So, testing the device in these conditions will be done too. In addition, the device does not have a backup system in case if something goes wrong, such as, the detection does not occur. Therefore, I would consider this as a future development. Due to COVID-19 the testing was limited using actual traffic light so it will be done when it is over.