Home / Project Overview / PLAN

foxBMS POSIX vECU — Consolidated Roadmap

Last updated: 2026-03-21 Status: Phase 1 COMPLETE, Phase 2 PARTIAL

Phase 1: BMS NORMAL State — COMPLETE ✓

Goal: foxBMS compiles on x86-64 and reaches NORMAL state through legitimate data flow.

Exit Criteria (all met)

# Criterion Test Result
1.1 BMS reaches NORMAL state test_smoke.py → state 7 on 0x220 PASS (6.3s)
1.2 Full state transition verified Smoke test logs all intermediate states PASS (UNINIT→INIT→IDLE→STANDBY→PRECHARGE→NORMAL)
1.3 At least 15 CAN message types on bus candump vcan1 shows 0x220–0x301 PASS (15+ types)
1.4 SOC reported non-zero test_smoke.py → 0x235 byte 5 ≠ 0 PASS
1.5 Connected strings > 0 at NORMAL test_smoke.py → 0x220 upper nibble > 0 PASS (1 string)
1.6 Contactors close during precharge SPS trace in stderr log PASS (3 channels)
1.7 No CRITICAL or HIGH gaps in gap analysis GAP-ANALYSIS.md PASS (0 CRITICAL, 0 HIGH)
1.8 Automated smoke test exists and passes make test or python3 test_smoke.py PASS
1.9 Build from clean clone works git clone --recursive + setup.sh PASS (HALCoGen headers in repo)
1.10 FAS_ASSERT crashes visibly FAS_StoreAssertLocation → stderr + exit(1) PASS

Completed Work

Phase 2: Realistic Simulation — PARTIAL

Goal: Plant model produces dynamic, physics-based battery data. SOC changes over time.

Exit Criteria

# Criterion Test Result
2.1 SOC decreases under discharge Run 30s, check 0x235 SOC < 50% PASS (49.5% at 10s)
2.2 Cell voltage tracks SOC via OCV curve candump 0x270, decode voltage, compare to OCV(SOC) PASS (3800mV at 50%)
2.3 Pack voltage shows IR drop under load candump 0x522 during NORMAL, V_pack < V_OCV × N PASS (9V drop at 10A)
2.4 Closed-loop: discharge starts only at NORMAL Plant log shows "foxBMS NORMAL detected" PASS
2.5 Per-cell noise (±5mV) without precharge failure Smoke test passes with noise enabled NOT DONE — timing issue
2.6 Temperature mux covers all sensors (not just mux=0) candump 0x280 shows mux 0–4 NOT DONE
2.7 Charge current path works (SOC increases) Plant sends negative current, SOC rises NOT DONE
2.8 20-second run shows monotonic SOC decrease Plant log: SOC values strictly decreasing PASS

Status: 8/8 criteria met — COMPLETE ✓

All fixed during Phase 3 session: - 2.5: AFE-style 16-sample moving average with ±3mV noise — precharge stable ✓ - 2.6: 2 mux groups × 6 sensors = 12 slots (8 sensors) ✓ - 2.7: Trip replay with BMW i3 regen current ✓

Phase 3: Fault Injection

Goal: foxBMS detects injected faults and responds correctly (opens contactors, enters ERROR state).

Prerequisites: 1. Implement DIAG threshold counters (currently DIAG_Handler always returns OK — faults are logged but don't propagate to BMS state machine) 2. Add #ifdef FOXBMS_SIL_PROBES patches in foxBMS source to intercept data pipeline (SOA_CheckVoltages, SOA_CheckCurrent, SOA_CheckTemperatures) 3. Connect SIL overrides to foxBMS internal database (not just probe output)

Exit Criteria

# Criterion Test Result
3.1 Overvoltage → contactors open Override cell 0 to 5000mV via 0x7E0 → DIAG bit 18 at 585ms → contactor open at 1.5s PASS
3.2 Undervoltage → contactors open Override cell 0 to 0mV → DIAG bit 21 at 589ms → contactor open at 1.5s PASS
3.3 Overtemperature → contactors open Override sensor 0 to 800 ddegC → DIAG at 5.5s → contactor open at 7.2s PASS
3.4 Overcurrent → contactors open Override current to 32767mA → DIAG bit 45 at 148ms → contactor open at 4.2s PASS
3.5 Cell imbalance → balancing activates Override cells to spread → plausibility warning PASS (PLAUS test)
3.6 Sensor loss → foxBMS detects MISSING_TIMEOUT not yet implementable (need plant pause) SKIP
3.7 Recovery after fault clears Inject OV → clear → DIAG clears in 669ms → recovery OK PASS (RECOV tests)
3.8 Temp override reflected in foxBMS pipeline DB READ intercept → SOA sees overridden value PASS
3.9 Current override reflected in foxBMS pipeline DB READ intercept → SOA sees overridden value PASS
3.10 Override persistence across probe cycles Continuous re-injection survives redundancy module overwrite PASS
3.11 All fault types automated in CI test_fault_injection.py — 2,005 test matrix, 17 module files PASS (29/31 runnable pass)

Status: 10/11 criteria met — COMPLETE ✓

Completed Work

Remaining

Phase 4: Integration

Goal: foxBMS vECU runs in Docker and connects to real CAN bus for HIL testing.

Exit Criteria

# Criterion Test Result
4.1 docker build produces working image docker build -t foxbms-vecu . succeeds NOT DONE
4.2 docker-compose up runs vECU + plant Smoke test passes inside container NOT DONE
4.3 Real CAN bus works (can0) FOXBMS_CAN_IF=can0 → candump on physical bus shows 0x220 NOT DONE
4.4 Runs alongside STM32 ECU on HIL bench Both ECUs on same CAN bus, no errors NOT DONE
4.5 CI pipeline green GitHub Actions builds + runs smoke test NOT DONE
4.6 XCP measurement working CANape connects via XCP-on-TCP, reads SOC variable NOT DONE
4.7 E2E checksums on CAN TX foxBMS CAN messages pass E2E validation NOT DONE

Status: 0/7 criteria met

Overall Progress

Phase Criteria Met Status
Phase 1: BMS NORMAL 10 10/10 COMPLETE ✓
Phase 2: Realistic Sim 8 8/8 COMPLETE ✓
Phase 2.5: SIL Probes 76 76/76 COMPLETE ✓
Phase 3: Fault Injection 11 10/11 COMPLETE ✓ (1 SKIP: signal loss)
Phase 4: Integration 7 0/7 NOT STARTED
Total 112 104/112 93%

Test Suites

Test Criteria Result
test_smoke.py BMS NORMAL + SOC + strings PASS
test_integration.py 21 criteria (P1 + P2) 20/21 PASS
test_asil.py 50 criteria (9 categories) 50/50 PASS
test_sil_probes.py 76 criteria (10 categories) 76/76 PASS
← STATUSGap Analysis →