Autonomous Vehicles

What we do


We develop optimal solutions for Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicles (AV). Our team has acquired unique expertise in computer vision algorithms, mathematical models for deep machine learning, data set conditioning and cloud infrastructures. These transverse skills allow us to provide the right solutions using the right technology that fit the constraints of real time and embedded applications.

We have in-depth knowledge of embedded platform solutions, as well as vision, lidar and radar technologies.  This allows us to achieve increased performance at low power consumption when using dedicated HW on-chip accelerators to integrate and optimize algorithms on target platforms.

As an example, we implemented a pedestrian detection function based on aggregated channel feature (ACF) on an automotive embedded platform using the hardware accelerators available. This resulted in a 12 fold runtime improvement over the same implementation on an Arm® Cortex® -A53 processor.

Our professional services allow our worldwide customer base to benefit from our advanced R&D expertise in this field.

eSoftThings is a member of the Renesas R-Car™ consortium, specialised in the Renesas autonomy™ Platform.

Our products


We build and condition datasets, create mathematical models to implement algorithm IP blocks for autonomous driving applications. These algorithms are optimized to address real-time and embedded constraints in order to meet the performances required by our customers on different platforms.

We also have developed eCube, our embedded ADAS Electronic Control Unit (ECU) that our R&D teams use internally to accelerate prototyping of advanced ADAS use cases. More information available here on how you can benefit accelerating time to market prototyping your ADAS use cases by adopting eCube in your teams. 

Professional services


We benchmark platforms to help our customers select the right solution to meet the performance requirements of their use cases.

Our team integrates and optimizes algorithms on embedded Platform solutions by utilizing the Platform dedicated on-chip accelerators to provide high performance at low power consumption.

We deliver high quality training on embedded Platforms for ADAS.

Use cases

Surround view

Front view

Driver monitoring

• Detection and classification of the environment

• Semantic analysis and cognitive analysis

• Sensor fusion