CAR DETECTION WITH OBSTANEX ON A REGULAR COMPUTER ARCHITECTURE : LOW DEPLOYMENT COST, LOW ENERGY CONSUMPTION, LESS HEAT, etc.
Here is a snapshot of car detection using RoadNex.
The big differenciation of ObstaNex is that is runs on a regular computer architecture (on a smartphone for instance) : no need for a heavy computing system. It is then much cheaper for mass volume deployment (new cars and aftermarket), because heavy computing architectures bring computing speed, but also high deployment cost, Energy consumption, heat, etc … not that good for onboard systems.
FREE SPACE IN THE DIRECTION OF THE WHEEL ON SHORT DISTANCE (URBAN DRIVING APPLICATION) : ON A REGULAR COMPUTER ARCHITECTURE IN REAL TIME
RoadNex Short (free space detection in front of the car) runs on regular computer architectures (even on a smartphone). This module is made for fast sensor fusion with lidar and radar.
It works even on dusty roads, stones (image below), cobblestones, etc …
RoadNex brings interpretation (drivable surface), telemeter (radar; lidar, …) brings measurement precision (in mm).
No need for a big computer (it means deployment cost reduction).
This disrupts some electronics architectures big firms that try to convince car manufacturers to put their computers Inside cars, but they do not bring only computing efficiency (they do), they also bring additional cost, weight, heat, integration room need, etc …
RoadNex runs on a regular ARM chip (for instance) and may be the next generation solution.
The next generation autonomous POD (Shuttle) MILLA made by ISFM uses RoadNex and will be shown at CES Las Vegas in Jan 2019.
Come to see it.
ANTICIPATION VS NINJA REFLEXES FOR ADAS AND AUTONOMOUS DRIVING : IMPACT ON ACCIDENT
Automotive industry is currently integrating into vehicles high level automations systems : automatic emergency braking, line keeping, etc … Those systems are complex : complex to do, complex to integrate together (as A system of systems), complex to validate.
Impact on accident of those complex features is unfortunately not that big. Indeed, accident is rare (1 accident per 70 000 to 100 000 km in OCDE countries) and the tree of risk situations (see image) branches that need “ninja reflexes” do not represent that much cases …
That is why NEXYAD proposes SafetyNex as an anticipation system that copes with the problem of “never being in emergency situation”. Of course, SafetyNex is not perfect (as every module) and it is important to keep emergency modules in the loop for a valuable collaboration.
If you have a look on the scheme below, you can notice that accident rate reduction brought by SafetyNex (anticipation effect) is much higher than ninja reflexes modules.
NEXYAD is currently integrating modules including SafetyNex into autonomous véhicles … to be continued.
NEXYAD Automotive & Transportation Newsletter #22, April 17th, 2018
4 disruptive AI algorithms for automotive mobility by NEXYAD
– SafetyNex episode 4 : Driving Risk Assessment for Automotive (Driving Assistant, ADAS, Autonomous Driving)
– CNEJITA Seminar on Artificial Intelligence: who will be responsible ?
– SafetyNex : driving robot maybe will mitigate human errors, but first they have to imitate good drivers
– “Theory of Water Flush” and Impact on the Prevention of Accidents for Autonomous Vehicles
– 4 disruptive AI algorithms for automotive mobility