Battery state-of-health as a functional safety variable: an ISO 26262-aligned AI framework for electric vehicle ADAS power integrity

A critical design blind spot persists in contemporary electric vehicle (EV) safety engineering: ISO 26262 functional-safety frameworks generally treat battery State-of-Health (SoH) as a passive maintenance metric rather than a live, mode-governing safety variable. This paper proposes a five-layer theoretical co-design framework that integrates SoH estimation, power-margin monitoring, functional-safety decision logic, actuation, and lifecycle management. A qualitative Fault Tree Analysis and preliminary Hazard Analysis and Risk Assessment are used to examine the representative BMS–ADAS power-interface hazard. The electrical analysis introduces a bounded DC/DC converter model and distinguishes two explicit converter-control scenarios: regulated operation with no active hard limit and an adverse 5.15 A hard-input-current-limit boundary condition. For the stated 851 W transient load, both scenarios retain positive converter margin; failed shedding of a 551 W comfort load is therefore the dominant source of power-margin compression, while the isolated SoH contribution is 0 W under regulated operation and approximately 40 W under the adverse hard-limit condition. A separate illustrative LiDAR subsystem-allocation model yields a beta-derived additional stopping-distance estimate of approximately 0.84–1.40 m, while an independently defined 100 ms recovery-delay scenario yields 1.67 m. These estimates are analytical scenario results rather than device-validated predictions. Platform-specific converter, power-allocation, sensor, timing, and hardware-in-the-loop validation is identified as the essential next steps.