NEXYAD is Deploying AI for Road Safety
NEXYAD Automotive & Transportation Newsletter #31, April 5th, 2021
Headlines :
– OASIS Project with MillaGroup, YoGoKo and Nexyad
– SafetyNex AI Technology for Driving Coach and Predictive ACC
– Intelligent & Predictive ACC with SafetyNex
– Road Safety Digital MAPS using Risk Data from SafetyNex
Hello, Despite the economic situation due to Covid-19, Nexyad is on a ramp up in this beginning of the year. We continue pushing innovation for our customers and partners. We Save Lives |
“We have integrated NEXYAD technology into our smartphone App BRIGHTMILE for insurers and fleets and our App reduces accident rate by 25-35% through its worldwide deployment in more than 50 countries. »
CEO of NEXYAD (left) and CEO of BRIGHTMILE (right): a handshake to save lives |
|
* * * * *
MillaGroup, Yogoko and Nexyad launch OASIS project with driver assistance solution for everyone French excellence for a decisive step towards road safety autonomous vehicles. Meudon, February 2021 – MillaGroup, Nexyad and YoGoKo launch the project in partnership OASIS (cOoperative Assistance System In Safe road), which announces new technological advances around cooperative driving assistance systems for all. The OASIS project proposes to develop and integrate new enhanced driving assistance features, based on an innovative data fusion solution. The data comes from V2X communications, driving risk assessment, and predictive collision estimation. Although the number of accidents decreases each year in France, more than a third of fatal accidents are linked to a collision. Today, cooperative ADAS solutions integrate data related to the human factor and driver inattention. The implementation of these driving aids (ADAS) takes into account the concept of risk. In driving, 75% of the risk is linked to inappropriate driver behavior, linked to road and urban infrastructure. The technology developed within the partnership is based on Nexyad’s expertise and knowledge of road risks. Safetynex is a data aggregation software module that detects 20 times per second any lack of caution (also called risk of driving) in driving behavior in view of the road context. Integrated into a real-time on-board AI, the solution makes the vehicle smarter and teaches it to anticipate certain situations, such as by slowing down preventively. Combining the Y-CONNECT solution with Y-SMART communication technology from YoGoKo increases the perception of the vehicle’s environment. YoGoKo combines V2X communications between vehicles and with the road infrastructure with the collision risk calculations related to road behavior offered by Nexyad. |
The OASIS project offers a global vision of road safety, deployed in marketed vehicles. This is experienced on the autonomous shuttles produced by MillaGroup. MillaGroup is a mobility provider offering on-demand and multimodal transport solutions (for people and goods) in rural and peri-urban areas. Since 2019, MILLA POD shuttles have been running on an open road every day at a speed of 30 km/h in autonomous mode. This combination of expertise in the area of risk management with Nexyad, connectivity with YoGoKo and vehicles deployment by MillaGroup will allow, among other things, an autonomous vehicle to drive thru an intersection with low visibility and priorities. Thus, the OASIS project brings new perspectives in response to the needs of improving road safety and increasing the perception of the vehicle’s environment. The system deployed, with risk traceability, takes into account the legal aspect and will make it possible to dispel the debate on liability in the event of an accident: – an assessment of driver behavior based on risk – improved predictive ACC and AD systems « aware » of the risk they are taking, and able to prove their level of caution at all times (in response to legal and insurance questions) – a new collision warning – a smarter predictive AEB Whether it is for the detection of lack of caution, dangerous situations, or emergency situations, road safety is integrated. This decompart- mentalised approach is completely innovative compared to the segmented approach by « use case » of the automobile. A system vision that must bring rapid progress in road safety, and on a larger scale respond to societal challenges. |
* * * * *
* * * * *
how to make ACC automatically follow rules of the road (slightly slow down before an intersection, etc.) without long and complex development? New developments are starting up everywhere to make ACCs smarter, which could, for example, automatically slow down slightly before an intersection, and then pick up the speed of initial cruise. The whole question is: when should the vehicle be made to slow down? crossing? turn ? school zone ? roundabout ? pedestrian zone ? pedestrian crossing? Traffic light ? stop sign ? give way ? etc. |
At every moment, SafetyNex detects any mismatch between driving behavior and driving context. Such a mismatch is interpreted as a lack of caution, and the degree of lack of caution is called « Driving risk »: this is the risk taken by the driver (human, human assisted by robots, or AD system). The higher the risk, the greater the number of emergency situations encountered by the vehicle. Conversely, the lower the risk, the lower the number of emergencies statistically encountered. Note : do not confuse the « driving risk » as defined above with criticality (probability of interpretation of trajectories, time to collision, etc.), the two concepts coexisting in road safety systems. For a given driving context, the driving risk depends ONLY on the speed profile of the vehicle approaching a potential singularity of infratructure. It then becomes clear that by regulating the speed we can keep the risk of driving below a maximum accepted value. SafetyNex is then used to define the setpoint speed value Vr (at each instant) which makes it possible to guarantee that we never exceed the maximum accepted. On the other hand, any current ACC already defines a speed setpoint Vo. By simply taking as the new speed setpoint the smaller of these two speeds, we obtain directly and without any further development, a vehicle that slows down before intersections (for example) with the « Good » speed profile which guarantees safe driving. This simple, validated, efficient, and quick to integrate system is currently being integrated by OEMs with support by NEXYAD. |
|
Contact us for more details. |
* * * * *
* * * * *
By Laurent MEILLAUD |
(Extract) « La technologie développée au sein du partenariat s’appuie sur l’expertise de Nexyad et sa connaissance des risques routiers. SafetyNex est un module logiciel d’agrégation de données qui détecte 20 fois par seconde tout mande de prudence (appelé aussi risque de conduite) dans le comportement de conduite et en fonction du contexte de circulation. Intégrée dans une IA temps réel embarquée, la solution rend le véhicule plus intelligent et lui apprend à anticiper certaines situations, comme par exemple en ralentissant de manière préventive. Pour rendre ce module encore plus performant, l’idée est donc d’y associer la technologie de communication Y-SMART de YoGoKo. Cela permet d’augmenter la perception de l’environnement du véhicule. YoGoKo associe les communications V2X réalisées entre véhicules et avec l’infrastructure routière aux calculs de risques de collisions liés au comportement routier proposés par Nexyad. Le projet OASIS est expérimenté sur les navettes autonomes produites par MillaGroup (un acteur dont j’ai déjà parlé sur ce blog). Depuis 2019, les navettes MILLAPOD circulent tous les jours sur route ouverte à la vitesse de 30km/h en mode autonome. |
|
We Save Lives