05Configuration simulator
Design your robot's sensor suite
Pick a platform, sensors and edge compute. The simulator estimates achievable autonomy, battery impact, cost and how the system copes with fog, darkness or no GPS — in real time.
Last reviewed: 2026-11Methodology & sources →
Simulated results
Level 5 is not a hardware question: it needs contextual learning and decision making not yet demonstrated in commercial deployments.
- Added power
- +63 W
- Battery impact
- −17 %
- Sensors + compute
- $3,299
- Base platform (ref.): $60,000
- Added weight
- 0.94 kg
- 7 % of payload
900 Wh · 300 W → 363 W
Environmental resilience
- FogLow
- No lightHigh
- Dense dustLow
- RainMedium
- No GPS / jammingHigh
Warnings
No warnings: consistent configuration.
Reference figures and where they come from
Each component is anchored to a representative commercial model: power and weight come from its datasheet (level B, or C when from a reseller or technical forum). Costs are ranges; the midpoint drives the totals and is usually an editorial estimate (D).
| Component | Reference model | Cost | Power | Weight | Evidence (spec · cost) | Source |
|---|---|---|---|---|---|---|
| Solid-state LiDAR | Livox Mid-360 | 700–$1,100 | 6.5 W | 0.265 kg | BC | [Livox][Robotics Center] |
| 360° mechanical LiDAR | Ouster OS1 (Rev7) | 6,000–$12,000 | 17 W | 0.45 kg | BD | [Ouster, 2023] |
| Radiometric thermal camera | Teledyne FLIR Boson 640 | 3,000–$7,000 | 1 W | 0.1 kg | BD | [Teledyne FLIR] |
| Stereo / VIO module | Intel RealSense D435i | 300–$500 | 2 W | 0.072 kg | BD | [RealSense] |
| RGB-D camera | Intel RealSense D435i | 300–$500 | 2 W | 0.072 kg | BD | [RealSense] |
| mmWave radar | TI IWR6843 (60–64 GHz) | 200–$1,000 | 2.5 W | 0.1 kg | CD | [Texas Instruments, 2025][TI E2E] |
| RTK GNSS | u-blox ZED-F9P | 300–$800 | 1 W | 0.15 kg | CD | [Mouser Electronics] |
| Ultrasonic ring | Generic | 100–$400 | 2 W | 0.15 kg | DD | — |
| Multi-gas sensor | Generic | 1,500–$5,000 | 3 W | 0.3 kg | DD | — |
| PTZ camera, 30× optical zoom | Generic | 2,000–$6,000 | 12 W | 0.9 kg | DD | — |
| Jetson Orin Nano Super | 67 TOPS · 7–25 W | $249 | ~15 W | 0.15 kg | BB | [NVIDIA Marketplace][CNX Software, 2024] |
| Jetson Orin NX 16GB | 157 TOPS · 10–40 W | $699 | ~25 W | 0.25 kg | BD | [NVIDIA] |
| Jetson AGX Orin 64GB | 275 TOPS · 15–60 W | $1,999 | ~40 W | 0.6 kg | BC | [NVIDIA][Hackster.io] |
| Jetson AGX Thor | 2070 FP4 TFLOPS · 40–130 W | $3,499 | ~90 W | 1.2 kg | CC | [Hackster.io] |
Base platforms
Editorial archetypes per family (level D) used to compute battery runtime; they do not represent a specific model.
| Battery | Base power | |
|---|---|---|
| Humanoid | 2000 Wh | 500 W |
| Quadruped | 900 Wh | 300 W |
| Ground vehicle | 2500 Wh | 300 W |
| Drone / aerial robot | 530 Wh | 560 W |
D· Editorial estimate
How the autonomy level is computed
- Level 1: IMU stabilization, included with every base platform.
- Level 2: localization + obstacle detection + compute ≥ 67 TOPS.
- Level 3: 3D perception + two localization sources + compute ≥ 150 TOPS.
- Level 4 (bounded ODD): LiDAR + GNSS-free localization + two obstacle modalities (one LiDAR or radar) + compute ≥ 250 TOPS.
- Added weight penalizes power by family (drones are the most sensitive).
- Thor is compared via an approximate TOPS equivalent; its published figure is FP4.
Model limits
- The rules are an editorial heuristic (level D): they count sensor modalities and compute capacity, not software performance.
- TOPS figures are the manufacturer's best-mode headline (67 sparse INT8 TOPS on Orin Nano Super = 33 dense), so they are not comparable across generations.
- Power draws are steady-state typical; LiDARs can draw up to twice as much at startup or in the cold (Mid-360 up to 14 W; OS1 ~28 W at −40 °C).
- Level 4 additionally requires validation in a bounded operational design domain (ODD); no hardware combination guarantees it.
Orientative simulation. Real autonomy depends mostly on software, data and validation in the target environment.