Design and Simulation of an Event Driven Sensorless IoT System for Automotive Battery Monitoring and Fault Detection

In the automobile industry, lead acid battery failure continues to be a major factor in vehicle immobility. In traditional monitoring systems, high inactive current draw, which adds to parasitic drain, or a lack of remote accessibility for real time user notification are common problems. A unique, event driven Internet of Things (IoT) framework for real time vehicle power diagnostics is proposed in this research. The sensorless architecture that uses straightforward resistor divider network to transmit voltage data straight into Analog to Digital Converter (ADC) of the microcontroller, is a significant innovation since it prevents the need for external sensors. Also, this suggested framework makes use of a state machine that alternates between an active sensing mode and a deep sleep period that is only activated by abnormal voltage occurrences. The proposed design validated through an electrical system and supervisory state logic modelled using MATLAB/Simulink, where Python serving as a secondary simulation and communication layer. Five main operating conditions are distinguished by the monitoring algorithm: parked, engine cranking, regular operation, alternator failure, and critical shutdown. The automotive voltage is scaled to an ADC compatible range using a 10 kΩ/2.7 kΩ resistor divider network, and signal filtering is provided by a 0.1 µF capacitor. With voltage settings of 12.60 V during parked operation, 9.80 V during engine cranking, 14.20 V during normal charging, 11.74 V during simulated alternator failure, and 10.30 V at critical shutdown, the simulation replicated the desired vehicle operating sequence. 1000 separate Monte Carlo runs with additive Gaussian measurement noise made up the stochastic validation. Relative mean voltage errors for the statistically assessed parked, alternator failure, and shutdown states were 0.0016%, 0.0009%, and 0.0029%, respectively, whereas the empirical detection success rate under the simulated conditions was 100%. The findings show that the suggested state machine architecture can reduce the need for ongoing high-power monitoring while differentiating between sustained abnormal voltage circumstances and normal beginning transients. As a result, the system offers a useful basis for controlled automotive validation and future ESP32 based physical implementation.