Problem statement
As the ecosystem of micro-mobility grows, ensuring the safety of eBike riders through V2X (Vehicle-to-Everything) technology is becoming increasingly important. A key challenge is bringing sophisticated hazard detections such as identifying approaching vehicles—to the constrained environment of eBike embedded system.
Current V2X solutions often demand significant computational power. To make these safety features viable, we need to integrate and optimize existing software solutions to run efficiently on low-power, real-time operating systems. The challenge lies in processing incoming V2X data and triggering real-time alerts on a bike display without oversubscribing the limited system resources.
Proposed solution
This thesis aims to develop a functional prototype of a V2X hazard detection system. The system will be two parts, one being an visualization device connected to the eBike system and the other being a specialized V2X hardware based on components from SpectrumFiftyNine. Both are based on Zephyr OS and these systems are to collaborate to enable the functional prototype.
The focus is on the "receiving" end of the V2X chain: processing environmental data to detect threats (e.g., cars on a collision course) and notifying the rider.
Implementation and evaluation:
-
Interopability: Ensure the systems communicate effectively, either via CanFD or via BLE.
-
Logic Development: Identify and notify of relevant objects (cars) based on incoming data streams.
-
Optimization: Ensure the system meets real-time requirements while operating within the performance constraints of an embedded microcontroller (focusing strictly on detection, not transmission).
-
Prototyping: Develop a functional "proof-of-concept" where warnings are successfully triggered on a bike display.
-
Benchmarking: Measure latency, CPU usage, and memory footprint to evaluate the feasibility of the solution for production.