Range-Finding Rear-Collision Accident Prediction and Warning For Motorcycles with Integrated Signal Vest
For the particular project at hand, a small amount of design work was completed. The design processes completed was the selection of a display method and sensor type. There was also the beginnings of part selection and theory design.
For the rear-mounted display, a system that shows simple pictures and would run on the motorcycles power supply. The simple pictures are on the order of a left and right arrow and an exclamation mark. The arrows are for the left and right turn signals while the exclamation mark is for the warning system. The first thought for the rear display was an LED monitor screen. The screen would be fairly bright and could easily display the pictures in mind. It would also allow for a design in mobile software to allow users to display their own pictures if they wanted to. The problem with this plan is the lack of flexibility of the screen and the power supply demands. After the screen, an array of individually addressable RGB LED's was considered. This method would allow for flexibility with the number of LED's in the grid and thus the power consumption. The individual addressing specification is to allow the LED's to be different colors from one another. Compared to most RGB LED strings, individually addressable LED's displays more than one color (if the user wants). LED's also take up very little power, so they will be able to provide a large amount of lumens and take up very little power. This design decision was the first decision of the project.
The automotive industry has two main methods for environment detection for accident prediction. These two are LIDAR and image processing. Both of these options were looked into as potential sensor systems for the project. A prebuilt LIDAR is too expensive for the project and requires more computing power than is available. Also, since the project is only for rear sided accident prediction, a full-blown LIDAR system is over engineering the problem. A camera and image processing system is also out of the equation. The image processing requires computing power that is not available for automotive applications by non-industry personnel. For these reasons, a camera system was not chosen. Since the accident prediction proposed is a "functional" accident prediction, the sensor system does not need to have a multitude of features. Based off of LIDAR and range finders, the sensor type of a "Time of Flight" (ToF) was chosen. ToF is used in LIDAR systems to get the distance between the target and the subject. Range finders use the same technology for the same purpose. This type of sensor has been proven to work in automotive applications by the application in LIDAR systems, so it is know that they technology can be used in the project's application. The sensor measures the distance between the motorcycle and the car behind it. Taking these measurements in timed intervals allows for the calculation of the speed and acceleration of the following vehicle. With these pieces of information, the accident prediction can determine if an accident (Distance = 0) will occur in the next 5 seconds. This design decision is the second portion of prior work.