EMBEDDED SYSTEMS
Effectively implementing intelligence algorithms on embedded systems is a real challenge.
While meeting your requirements of accuracy, robustness and reliability, our parsimonious AI breaks down all barriers related to:
-
Bulk constraint
-
Energy consumption
-
Computional resources
-
Data transmission bandwidth
Adagos has developed a parsimonious AI, oriented to technical data.
Our technology is specifically adapted to embedded systems.
In particular, for low orbit satellites, our models are 1000 times smaller than what is typically use to control the antennas of the satellites.
Moreover, unlike existing solutions, the embedded software is completely free.
OUR SUCCESS STORIES
Satellites and Antennas
Car production
Car racing
Combustion optimization
Robotic factory applications
Autonomous driving
Satellites and Antennas
Embedded system within Satellites and Antennas
Partner objective
Adagos has developed a parsimonious AI, oriented to technical data. The goal is to enable the satellite to provide a stable position of the radiated energy on the earth as it orbits.
Why NeurEco
-
The embedded software is automatically created by NeurEco.
-
Development and deployment times are reduced significantly due to less data requirement.
Benefits to the client
-
In particular, for low orbit satellites, our models are 1000 times smaller than what is typically used to control the antennas of the satellites.
-
Unlike existing solutions, the embedded software is completely free.
Car production
Working with a leading Car Manufacturer
Partner objective
Car manufacturer would like an embedded software that will optimally control the car using low cost sensors.
Why NeurEco
-
The companies conventional modeling was not able to account for all of the dynamic inputs.
-
NeurEco’s dynamic module accurately and efficiently produced an embedded algorithm that was able to control the car more that the Car Manufacturers engineers thought was possible.
Benefits to the client
-
Development time reduction: the model has been created automatically with the default setting of NeurEco reducing the requirement for expensive sensors.
-
NeurEco reduced the energy consumption of the car.
-
The created model being small, it reduces the cost of the embedded equipment.
Autonomous driving car
Partner objective
Reduce the model size needed for controlling an autonomous car to allow the software to be embedded.
Why NeurEco
-
With its default settings NeurEco created a neural network 100 times smaller than the client's. The calculations were accelerated by a similar factor.
-
All NeurEco’s outputs are fully embeddable in the car.
Benefits to the client
-
Development time reduction: the model has been created automatically with the default setting of NeurEco.
-
As the created model is small, it reduces the cost of the embedded equipment.
-
Reduction in energy consumption of the car resulting in longer battery life and range.
Combustion engine optimization
Partner objective
Embedded control of a thermal engines combustion.
Why NeurEco
-
The NeurEco dynamic add-on is unique on the market.
-
NeurEco is able to model, in real time, all of the input data due to the efficient model size resulting in fast processing time and optimal engine control.
Benefits to the client
-
Development time reduction: the model has been created automatically with the default setting of NeurEco. Client estimated a reduction of 6 months to R&D time.
-
NeurEco reduced the fuel consumption of the engine aligned with the companies carbon objectives.
-
As the created model is small, it reduces the cost to embed, the speed of the model allows for fast changes to the engine and accuracy allows for extreme reliability.
Robotic application
Partner objective
Reduce the resources needed to embed software in a robotic production system.
Why NeurEco
-
Many companies struggle to accurately model dynamic motions. NeurEco has a unique dynamic modeling solution capable of doing this.
-
With its default settings NeurEco created a neural network 1000 times smaller than the client’s original.
-
The speed of the calculations were accelerated by a similar factor leading to the algorithm being able to implement any needed changes to the robot in a very short space of time.
Benefits to the client
-
Development time reduction: the model has been created automatically with the default setting of NeurEco.
-
As the created model very small, it dramatically reduced the cost to embed and allowed for easier model explanation.
-
Furthermore as the model was small it used a smaller amount of energy extending the autonomy of the robot.