Integrated CAN-Bus System: A Software-Defined Multi-Threaded Instrument Cluster and Open Observability Telemetry Architecture for Vehicle Performance Optimization
Abstract
The adoption of standalone Engine Control Units (ECUs) in motorsport and research vehicles introduces severe interoperability challenges with OEM electronic architectures, often rendering factory instrument clusters inoperative. Existing commercial displays address this but feature high costs, closed ecosystems, and limited extensibility. This paper presents a low-cost, software-defined automotive instrument cluster and telemetry platform built on a Raspberry Pi 4, bridging heterogeneous vehicle subsystems without modifying ECU firmware. The architecture unifies high-rate CAN bus acquisition, analog digitization, discrete chassis signal monitoring, and GNSS telemetry. To evaluate network expansion, an auxiliary Exhaust Gas Temperature (EGT) loop using a MAX31855 amplifier was integrated to broadcast thermal data over CAN to a Link G4+ ECU. A priority-scheduled multi-threaded software model ensures bounded application-level acquisition alongside graphical visualization and an onboard time-series database (TSDB) with an embedded Grafana observability framework. Experimental validation on a cross-manufacturer testbed (Mazda RX-8 with a Nissan SR20DET engine and Link G4+ ECU) demonstrated stable operation during long-duration testing. Stress-testing up to 1000 frames/s with no frame loss was observed. The results prove the platform provides a highly flexible, extensible, and economically accessible open-source alternative to proprietary motorsport instrumentation.
