Introduction
A systems engineering exercise applying classical function-to-form synthesis to decompose an electric vehicle's performance requirements into a physical subsystem architecture, then comparing the result against the original program specification tree.
The source document is a Vehicle Technical Specification for a battery electric vehicle intended for commercial sale to the general public. The specification defines performance requirements across six functional areas: transportation, safety, comfort, enjoyment, energy management, and utility. It does not prescribe a physical architecture.
The exercise asked Claude Cowork to perform function-to-form synthesis: reading the specification, identifying all performance requirements, and mapping them to physical subsystems that together form a fully integrated vehicle design. The AI-generated architecture was then compared to the original program's specification tree (GSP3) to evaluate convergence and divergence in subsystem decomposition decisions.
Source Requirements
The Vehicle Technical Specification (74 pages) defines performance requirements in Clause 3.2.1 across six major areas: Transportation (3.2.1.1), Safety (3.2.1.2), Comfort (3.2.1.3), Enjoyment (3.2.1.4), Energy Management (3.2.1.5), and Utility (3.2.1.6).
Design Method
Function-to-form synthesis is the iterative process of mapping what a system must do (functions derived from requirements) to what physical components will be responsible for doing it (forms). The process is the core synthesis step in systems engineering, consistent with ISO/IEC/IEEE 15288.
Tool Used
The AI design was produced using Claude Cowork (Opus 4.6). Cowork read the scanned specification PDF, extracted requirements via OCR, conducted the synthesis autonomously, and produced a Word document with subsystem definitions, a requirements allocation matrix, and an architecture diagram.
Output
The synthesis produced: 12 named subsystems, 2 vehicle-level cross-cutting disciplines, a complete requirements allocation matrix (54 clause areas), interface definitions, and an architecture diagram — all documented in a formal report (Rev 2).
Systems Engineering Method
The classical four-step process for translating functional requirements into a physical architecture of subsystems and interfaces.
Function-to-form synthesis begins with a complete set of performance requirements and ends with a physical subsystem architecture. The process is iterative: initial allocations are refined as interface conflicts, cross-cutting concerns, and integration constraints are discovered. The key design heuristics are high cohesion (functions sharing state belong together), low coupling (loosely coupled functions should be separated), and alignment with organizational boundaries (subsystems should map to design teams, suppliers, and manufacturing processes).
All performance requirements from Clause 3.2.1 were extracted and catalogued across six functional areas: Transportation, Safety, Comfort, Enjoyment, Energy Management, and Utility — totaling 54 major clause areas.
Each top-level requirement area was decomposed into sub-functions at sufficient detail to expose distinct physical form assignments. For example, "Safety" decomposes into crashworthiness, restraints, visibility, and operational safeguards.
Each function was allocated to one or more physical subsystems using cohesion, coupling, and organizational alignment principles. A requirements allocation matrix records primary and contributing subsystem assignments for all 54 clause areas.
Data, energy, and control flows between subsystems were identified and documented. Critical interfaces (regen braking blend, torque vectoring, battery thermal management) were highlighted with derived interface requirements.
Requirements Summary
The specification organizes performance requirements into six functional areas, each addressing a distinct aspect of the vehicle's mission.
3.2.1.1 — Transportation
Acceleration (0–60 in target time), top speed, gradeability, range per charge, cruise control, braking distances, and ride/handling characteristics including steering effort and stability.
3.2.1.2 — Safety
Crashworthiness (frontal, side, rear impacts), airbag deployment, seat belt performance, visibility (mirrors, lighting, wipers), HV isolation, and operational safeguards (park brake, interlocks).
3.2.1.3 — Comfort
HVAC heating/cooling capacity, defrost/defog performance, interior noise levels at speed, vibration modes, ride comfort, and seating ergonomics for driver and passenger.
3.2.1.4 — Enjoyment
Audio system (radio, speakers, amplifier), interior styling and trim quality, exterior appearance including paint, moldings, and lighting design.
3.2.1.5 — Energy Management
Battery capacity and charging (Level 1/2), regenerative braking energy recovery, accessory power budget, thermal management efficiency, and EMC performance (radiated/conducted emissions).
3.2.1.6 — Utility
Jack and tire change provisions, cargo capacity, towing/trailering, corrosion resistance targets, and service access for maintenance operations.
AI-Generated Architecture
Claude's synthesis produced 12 physical subsystems and identified two cross-cutting concerns that operate as vehicle-level disciplines rather than owned subsystems.
SS-01
Structural Performance
Load-bearing skeleton, crash energy management, body stiffness, HV battery enclosure isolation.
SS-02
Occupant Accommodation
Seating, IP and console, sidewall trim, headliner, carpet, insulation, steering wheel and column.
SS-03
Restraint Performance
Airbags (frontal/side), seat belts, pretensioners, head restraints, crash sensing and SIR electronics.
SS-04
Exterior Envelope
Body panels, doors, bumpers, exterior lighting, mirrors, wash/wipe, coating and moldings.
SS-05
Thermal & Climate Control
Heat-pump HVAC, battery thermal management, defrost/defog, cabin preconditioning, odor control.
SS-06
Braking Performance
Blended regenerative + friction brakes, ABS, park brake, pedal feel, brake thermal control module.
SS-07
Ride, Handling & Steering
Front/rear suspension, electric power steering, tires and wheels, chassis structures, stability control.
SS-08
Accessory Power
12V bus, DC-DC converter, auxiliary battery, power distribution module, wiring harness, load management.
SS-09
Entertainment
Radio, antenna, front/rear speakers and enclosures, audio amplifier, audio controls.
SS-10
Information & Controls
Digital cluster, PRND selector, cruise control, mode switches, warnings, door switches, steering column modules.
SS-11
Propulsion Performance
Dual motors, inverters, battery pack, power electronics bay, on-board charger, inductive charge module, accelerator pedal sensor.
SS-12
Security & Operational Safeguards
Immobilizer, key fob/entry, parking device, power windows, operational interlocks and safety logic.
VL-A
EMC Performance
Vehicle-level discipline: radiated/conducted emissions and susceptibility design rules flowing down to all subsystems.
VL-B
Acoustic Performance
Vehicle-level discipline: interior/exterior noise targets, NVH target cascade, squeaks and rattles, vibration modes.
Figure 1 — AI-Generated Subsystem Architecture (Rev 2): 12 subsystems arranged by functional tier, with color-coded critical, primary, and supporting interfaces. Red dashed bands represent vehicle-level disciplines.
Original Program Architecture
The original program's specification tree (GSP3, dated 1/31/92) defined 11 subsystems plus EMC and Acoustic Performance as vehicle-level disciplines outside the subsystem hierarchy.
1
Structural Performance
Front/center/rear structure, fixed glass, jack and tools. (STRUCT)
2
Occupant Accommodation
IP and console, seats, sidewall trim, headliner/carpet/insulation, steering wheel and column. (ACCOMM)
3
Restraint
Active belts, driver and passenger SIR modules, SIR electrical system. (RESTRT)
4
Exterior Envelope
Front/rear end panels, door structures, bumper systems, moldings/rockers/sails, exterior lighting, coating, mirrors, wash/wipe. (EXTAPP)
5
Thermal and Climate Control
HPVM, refrigeration, solar air exhaust, air ducts, HTCM, CRFM, coolant distribution, power handling devices, auxiliary heater. (CLIMAT)
6
Braking Performance
Pedal module, master cylinder, brake modulator assembly, pipes/hoses, front/rear corner assemblies, BTCM, park brake control. (BRKING)
7
Ride, Handling and Steering
Chassis structures, front/rear suspension, drive axles, motor mounts, steering subsystem, tires and wheels. (RIDEHN)
8
Accessory Power
Auxiliary power supply, auxiliary battery, power distribution wiring. (ACCPOW)
9
Entertainment
Antenna, radio, front/rear speakers and enclosures. (ENTRTN)
10
Information and Controls
Driver information, CCU, steering column modules, door switches, external audible warnings. (INFORM)
11
Propulsion Performance
Battery pack, power electronics bay, drive unit, inductive charge module, accelerator pedal sensor. (PROPUL)
SSTS 13
EMC Performance
Vehicle-level: radiated/conducted emissions and susceptibility. Design rules flow down to all subsystems.
SSTS 15
Acoustic Performance
Vehicle-level: interior/exterior noise, NVH targets, vibration modes. Target cascade to all subsystems.
Figure 2 — GSP3 Specification Tree: Original program architecture with 11 subsystems hanging from a single bus bar, plus EMC and Acoustic Performance as vehicle-level disciplines. All subsystems trace to the Vehicle Technical Specification.
“The test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function.”
— F. Scott Fitzgerald (adapted for systems engineering)The AI-generated architecture (Rev 2, 12 subsystems) was refined after comparison with GSP3. The two designs converge strongly on 10 of 11 GSP3 subsystems — the AI independently arrived at nearly identical boundaries for structural, occupant accommodation, braking, ride/handling/steering, climate, exterior, entertainment, information/controls, accessory power, and propulsion.
The principal divergence is the AI's addition of SS-12 (Security and Operational Safeguards) as a separate subsystem, whereas GSP3 distributes those functions across Information and Controls and other subsystems. This reflects a modern design sensibility where security is an increasingly distinct engineering concern.
Comparison Summary
Strong Convergence
10 of 11 GSP3 subsystems have direct one-to-one matches in the AI design, with nearly identical functional scope and component allocation.
Vehicle-Level Agreement
Both architectures elevate EMC and Acoustic Performance to vehicle-level disciplines outside the subsystem hierarchy — recognizing these as cross-cutting design concerns.
AI Addition
SS-12 (Security and Operational Safeguards) is unique to the AI design. GSP3 distributes key fob, immobilizer, and interlock functions across existing subsystems.
Iterative Refinement
The AI's initial Rev 1 (10 subsystems) over-split ride/handling/steering and under-differentiated interior vs. exterior. Rev 2 corrected these after reviewing GSP3.
Detailed Comparison
How each GSP3 subsystem maps to the AI-generated architecture, with alignment assessment.
| GSP3 Subsystem | AI Subsystem (Rev 2) | Alignment | Notes |
|---|---|---|---|
| 1. Structural Performance | SS-01 Structural Performance | STRONG MATCH | Identical scope: body shell, crash energy management, stiffness, HV isolation. |
| 2. Occupant Accommodation | SS-02 Occupant Accommodation | STRONG MATCH | Same component set: IP, seats, trim, insulation, steering wheel. |
| 3. Restraint | SS-03 Restraint Performance | STRONG MATCH | Both isolate airbags, belts, and SIR as a distinct subsystem. |
| 4. Exterior Envelope | SS-04 Exterior Envelope | STRONG MATCH | Panels, lighting, mirrors, wash/wipe, coating — identical boundaries. |
| 5. Thermal & Climate Control | SS-05 Thermal & Climate Control | STRONG MATCH | AI adds battery thermal management explicitly; GSP3 includes it via HTCM/CRFM modules. |
| 6. Braking Performance | SS-06 Braking Performance | STRONG MATCH | AI emphasizes regen-friction blend; GSP3 lists same hardware components. |
| 7. Ride, Handling & Steering | SS-07 Ride, Handling & Steering | STRONG MATCH | Merged in Rev 2 to match GSP3. Rev 1 had incorrectly split this into two subsystems. |
| 8. Accessory Power | SS-08 Accessory Power | STRONG MATCH | 12V bus, DC-DC, auxiliary battery, power distribution. |
| 9. Entertainment | SS-09 Entertainment | STRONG MATCH | Separated in Rev 2. Rev 1 had incorrectly merged entertainment into information/controls. |
| 10. Information & Controls | SS-10 Information & Controls | PARTIAL MATCH | AI separates security functions into SS-12; GSP3 keeps door switches and warnings here. |
| 11. Propulsion Performance | SS-11 Propulsion Performance | STRONG MATCH | Battery pack, power electronics, drive unit, charger, accelerator pedal sensor. |
| — | SS-12 Security & Oper. Safeguards | AI ADDITION | No GSP3 equivalent. Immobilizer, key fob, interlocks collected from other subsystems. |
| EMC Performance (SSTS 13) | VL-A EMC Performance | STRONG MATCH | Both treat EMC as a vehicle-level discipline, not an owned subsystem. |
| Acoustic Performance (SSTS 15) | VL-B Acoustic Performance | STRONG MATCH | Both treat acoustics as a vehicle-level NVH target cascade. |
Observations
Convergence
Working only from the requirements text, the AI arrived at nearly the same subsystem boundaries as the original program team. This suggests that well-structured requirements strongly constrain the design space, and that function-to-form synthesis is a learnable, reproducible process.
Iteration
The AI's Rev 1 made characteristic errors: over-splitting ride/handling/steering and merging entertainment into controls. Comparing against GSP3 prompted corrections in Rev 2, demonstrating the value of human-AI architecture review cycles.
Modern Sensibility
The AI's addition of SS-12 (Security) reflects a contemporary engineering perspective where security, cybersecurity, and operational safeguards are increasingly treated as first-class architectural concerns rather than distributed across other subsystems.
Method
This exercise demonstrates that AI can perform classical systems engineering synthesis from a specification document, producing subsystem architectures, requirements allocation matrices, and interface definitions that are directly comparable to human expert output.
Interested in function-to-form synthesis, EV architecture, or AI-assisted engineering design? Get in touch to discuss.
✉️ percivall@ieee.org