Sense. Sync. Predict. Perform.
S2P2™ Edge is the shared embedded software platform for Ride Pilot™ and RDAS™. It synchronizes perception, human and vehicle data, predicts risk on-device, and supports safety-arbitrated vehicle control through one deterministic loop.
Four stages. One real-time loop.
Sense
Cameras, radar, IMU, and vehicle signals build a continuous picture of the road, human, and machine state.
- —Camera + radar
- —IMU & vehicle bus data
- —All-weather, day & night
Sync
Every sensor stream is time-aligned and fused into a single world model with deterministic latency.
- —Hardware timestamping
- —Sensor fusion
- —CAN-FD & Ethernet IO
Predict
On-device ML forecasts trajectories, risk, and human intent using models tuned to each product and vehicle class.
- —Trajectory & TTC prediction
- —Rider and driver context
- —Edge-case models from the twin
Perform
Automotive-grade control translates decisions into bounded warnings or vehicle action, with product-specific safety arbitration.
- —Precision vehicle control
- —Display & mobile alerts
- —Hardware + software safety loops
One real-time platform for two robotic products.
Ride Pilot™
The robotic product for compatible CVT scooters and electric bikes and scooters. S2P2™ Edge runs its real-time intelligence.
Explore Ride Pilot™ →RDAS™
The Robotic Drive Assist System for cars and four-wheelers. S2P2™ Edge runs its real-time intelligence and control software.
Explore RDAS™ →S2P2™ Twin
The separate digital-twin platform that trains and validates Ride Pilot™ and RDAS™ models before deployment to S2P2™ Edge.
Explore S2P2™ Twin →Everything you need. Nothing you don't.
On-device inference
All safety-critical models run locally; the cloud is optional.
Signed & encrypted OTA
Model + firmware bundles are cryptographically verified end-to-end.
Deterministic bus IO
CAN-FD, Ethernet, and LVDS with bounded latency budgets.
Sensor abstraction
Drop-in support for stereo, mono, imaging radar, and thermal.
Telemetry & fleet ops
Optional aggregated metrics — no rider PII leaves the vehicle.
Safety case tooling
Traceability from requirement to scenario to signed artifact.
A thin, typed interface — nothing hidden.
Connect to a device, subscribe to perception events, and stream telemetry with a few lines of TypeScript.
import { RiderSession } from "@automata/sdk";
const session = await RiderSession.connect({
device: "automata-edge-01",
scopes: ["perception.events", "telemetry.aggregate"],
});
session.on("collision.warning", (evt) => {
// evt.ttc: number (seconds)
// evt.confidence: number (0..1)
// evt.object: { class, bearing, closingSpeed }
console.log(`TTC ${evt.ttc.toFixed(2)}s · ${evt.object.class}`);
});Rider safety needs its own physics engine.
S2P2™ Edge is configured to the dynamics and safety needs of each robotic product. For Ride Pilot™, it runs rider-aware intelligence on compatible CVT scooters and electric two-wheelers; for RDAS™, it supports four-wheeler perception, prediction, and control.
Two-wheeler-aware AI
Dynamics-aware perception and prediction tuned for supported CVT scooters and electric two-wheelers.
S2P2™ Twin
Train and validate Ride Pilot™ and RDAS™ models against repeatable safety scenarios before deployment.
Edge AI Module
On-device inference running locally with deterministic latency budgets.