Echodar offers a software‑only processing stack that converts raw automotive radar point clouds into high‑density, semantically enriched data for object detection, classification and SLAM. The platform combines AI‑augmented filtering with synthetic point injection to improve resolution and suppress clutter, and provides an open API compatible with ROS, AUTOSAR and other sensor‑fusion frameworks for deployment on ECU‑grade processors or cloud edge nodes.
Funding
Funding not disclosed
Founders
Product
Problem
Automotive radar sensors generate sparse point‑cloud data that often lack the resolution needed for reliable object detection, classification, and simultaneous localization and mapping (SLAM) in advanced driver‑assistance systems (ADAS) and autonomous driving. Existing hardware upgrades are costly and time‑consuming, while raw radar data is prone to clutter and weather‑induced noise, limiting perception performance. Consequently, OEMs and Tier‑1 suppliers must rely on additional sensors or extensive calibration to meet safety requirements.
Solution
Echodar delivers a software‑only processing stack that transforms raw radar point clouds into high‑density, semantically enriched representations without modifying the underlying hardware. The platform combines traditional signal‑processing filters with AI‑driven denoising and synthetic point injection to boost effective resolution and suppress clutter across all weather conditions. Processed data include precise range, velocity, object type, size, and trajectory information, enabling robust detection, classification, and SLAM pipelines. An open API allows seamless integration with existing sensor‑fusion frameworks, while the solution scales to any automotive radar architecture. By offloading computationally intensive steps to optimized on‑vehicle or cloud runtimes, Echodar reduces development cycles and lowers the total cost of perception stacks for autonomous driving.
Target Audience
The primary customers are automotive OEMs and Tier‑1 suppliers building ADAS and Level‑3/4 autonomous driving systems that require enhanced radar perception without hardware redesign. System integrators and software vendors developing sensor‑fusion or mapping solutions also benefit from the API‑first architecture.
Features
- AI‑augmented filtering pipeline that fuses deep‑learning denoising with physics‑based clutter rejection (speed, range, angle)
- Synthetic point injection engine that adds >0 % virtual points per frame, effectively increasing radar angular resolution to near‑camera levels
- Real‑time object classification and tracking output (vehicle, truck, pedestrian) with velocity and size attributes
- SLAM‑ready map generation module delivering high‑precision road‑boundary and obstacle layers for localization
- Vendor‑agnostic SDK and RESTful API supporting ROS, AUTOSAR, and proprietary sensor‑fusion stacks
- Configurable deployment on ECU‑grade processors or cloud edge nodes with deterministic latency guarantees
- End‑to‑end encryption and ISO‑26262 compliant data handling for safety‑critical applications