How do we plan a closed-track testing session?
- seaxaidevloop
- Aug 4
- 3 min read
Updated: Aug 14
Lately, we’ve been asked how we go from office simulations to a vehicle that operates entirely on its own. The short answer: simulation is a prerequisite, never the destination. Simulation gets us closer, but reality is always different, so we use it to reduce iterations and track time.
Each testing session is designed with measurable objectives, controlled scenarios, and safety above all else. In this run, we focused on making the ride more comfortable by adjusting speed in curves and driving for 30 minutes straight without any human intervention in dynamic control.
Session objectives
Comfort in curves: Dynamically adjust speed to keep lateral acceleration within a comfortable threshold. In practice, this means entering each curve at the right speed so passengers feel a smooth transition—no jerks or sudden body roll.
Continuous operation without intervention (30 min): Demonstrate that the vehicle can complete long sessions without human corrections to steering, brakes, or accelerator, maintaining stable trajectory tracking and a steering wheel free of nervous microcorrections (with no obstacles on the track).
How we tested it
Before hitting the track: We simulated the scenarios and defined success criteria.
On the track (closed circuit): We validated that both sensors and actuators responded properly and ran the test, which involved completing various obstacle-free maneuvers.
Safety: Active geofence, pre-run checklist, safe stop plan, and a safety operator in the driver’s seat.
What we observed during the test
In curves, the vehicle slowed down in advance, entered with controlled lateral acceleration, and maintained its line without “saw-tooth” patterns in the steering. On exit, it regained speed progressively, avoiding jerks. When control didn’t require sharp steering corrections, the onboard feeling was one of comfort and predictability: less body roll, smoother transitions, and no surprises for passengers.
During the long session (30 minutes without intervention), trajectory tracking remained stable: the car corrected small deviations slowly and smoothly, without quick micro-turns that would indicate “nervous” control. The steering pattern felt smooth and continuous, and the driving maintained a steady rhythm even in linked curves (left-right sequences).
Translated into sensations: the car didn’t brake aggressively before turning, didn’t “jolt” when entering or exiting curves, and didn’t vibrate on straights from unnecessary microcorrections. The result: a more comfortable and predictable experience, bringing us closer to our goal of adjusting speed to curvature without sacrificing efficiency.
Note: In simple terms, the system calculates the right speed based on road curvature to keep lateral acceleration “in the comfort zone.” This way, comfort is prioritized without losing efficiency.
What we took from this run
Validated the curvature-based adaptive speed approach to improve ride experience.
Confirmed that the platform can run for 30 minutes straight without dynamic intervention, with predictable behavior.
Next steps: Expand scenarios (tighter curves, uneven surfaces) and compare different control configurations to refine comfort and efficiency.
Quick glossary
Lateral acceleration: The “force” you feel to the side when going through a curve.
Trajectory tracking: How well the vehicle stays on the planned path.
Microcorrections: Small, rapid steering adjustments that, when accumulated, create a feeling of vibration.
Safety → Tests carried out on a closed circuit, with specific protocols and a safety operator. Not representative of open-road conditions.
Interested in following our progress?
Subscribe or contact us if you represent a municipality or operator and want to explore pilot projects in a rural environment.
Comments