Simulation as Essential Infrastructure During Lockdown
With lab access limited, simulation becomes even more critical - accelerating our investment in virtual engineering.
Silver lining of the pandemic: simulation investment that seemed "nice to have" is now essential. We can't access labs, but we can access simulated labs.
The Lockdown Acceleration
Before March 2020:
- Simulation supplemented physical testing
- Real data collection was primary
- Hardware prototyping happened in labs
After March 2020:
- Simulation must substitute for physical testing
- Synthetic data is often the only option
- Hardware simulation enables progress without labs
Virtual Engineering Expansion
We've expanded simulation beyond training data:
Sensor Prototyping
Before: Build prototype sensor, test, iterate. Now: Simulate sensor characteristics, predict performance, iterate in simulation.
Simulating:
- Optical path (ray tracing through lens systems)
- Sensor response (photon-to-electron conversion, noise models)
- Processing pipeline (depth estimation, feature extraction)
Result: Can evaluate sensor design changes in hours, not months.
System Integration
Before: Integrate components on bench, find issues, debug. Now: Simulate integrated system, predict issues, fix before hardware exists.
Simulating:
- Timing interactions between sensors
- Data flow bottlenecks
- Failure mode propagation
Performance Prediction
Before: Profile on target hardware. Now: Cycle-accurate simulation of target processors.
Not perfect, but good enough to catch major issues before silicon.
Infrastructure Investments
Accelerated investments:
- Render farm capacity: 3x increase in cloud GPU allocation
- Physics simulation: License expansion for optical simulation tools
- Asset library: More 3D environments, objects, materials
- Automation: Better job scheduling, monitoring, analysis
Limitations We're Hitting
Simulation isn't magic:
Fidelity ceiling: No matter how good the simulation, reality surprises.
Validation gap: How do you validate a simulation without access to the real thing?
Tacit knowledge: Some debugging intuition only comes from physical experience.
Human factors: Can't simulate the experience of wearing the device.
Hybrid Workflow
When we do get lab access (limited, scheduled):
- Prioritize validation: Use physical testing to validate simulation predictions
- Collect calibration data: Improve simulation fidelity with real measurements
- Edge case exploration: Investigate things simulation couldn't predict
Future State
Post-pandemic, we won't go back to pure physical development:
- Simulation-first for early design phases
- Physical validation for maturity milestones
- Continuous refinement of simulation accuracy
The pandemic forced an evolution we should have made anyway. Virtual engineering is now core to our process.