NEXYAD Automotive & Transportation Newsletter #32, June 8th, 2021
NEXYAD has developed a unique cognitive AI platform that aggregates all available sources on board a vehicle, in real time during the trip. In particular it aggregates : digital map, electronic horizon, GNSS, accel, jerk, speed, diagnostics at the output of ADAS sensors (camera, radar, lidar, …) with detection of pedestrians, cars, truck etc, driver monitoring, V2X hazard warnings. This data aggregation technology is called SafetyNex. In real time, and while driving, 20 times per second, SafetyNex calculates:
The 3 typical applications are to date:
. the Intelligent ACC, which modulates the speed of the vehicle, including in dense urban areas, in order to guarantee a given level of caution. Every time that driving risk could rise because unsafe acts are detected in driving behavior, the car modulates its speed to stay in the « green zone ». Again, mitigating risk (driving in a cautious way) minimizes occurrence of emergency situations that may lead to accident, and then minimizes occurrence of emergency quick reflexes of the car (e.g. AEB) that are not good for comfort.
This entering in the automotive market is a second step for NEXYAD that has worked for 20 years to design, develop, perfect, and industrialize this hybrid cognitive AI. Together we will prevent more traffic accidents and save more lives.
SafetyNex is the first and unique on-board hybrid AI platform that aggregates in real time all the data available on board (static as dynamic) in the vehicle and calculates 20 times per second a level of caution, a value of driving risk, to deliver alerts, the maximum speed not to be exceeded in order to remain cautious, given the driving context (here and now).
The « Safety Coach » application is a variation of SafetyNex integrated into the vehicle and which benefits from the available input data (automotive quality). HMIs can, of course, differ from one OEM to another. The common thread is that they are proactive safety coaches: coaching takes place during the trip, a few seconds before potential difficulties, and advises the driver to slow down (for example), instead of doing so after the trip.
Important: this does not prevent the application from offering a debriefing after the trip which is suitable for professional drivers, as well as for privates. In the image below, we can see the value of the risk of a trip (level of lack of caution). As our risk scale has been calibrated with road safety experts from 19 countries, 95% of global accidents happen between 90% and 100% risk taking.
Proactivity of the Safety Coach
The Safety Coach is proactive, he doesn’t just send summaries, he interacts in real time (HMI to be developed as part of an OEM application, one of the avenues being integration into the navigation system): when the driver crosses the value 90%, it is important to alert it. This alert (red) is not an emergency alert, it is an anticipatory alert ultimately expressing « you are being reckless ».
The driver then has time to take his/her foot off the accelerator (in general), and to return to safe driving, without stress and without loss of comfort. In doing so, it decreases the bad luck of finding itself in an emergency situation that would trigger safety ADAS (such as emergency braking, for example). You can see a yellow-orange line called « yellow alert » or « orange alert ». We recommend leaving the setting of this alert in the hand of the driver: if the driver feels reassured by « talkative » assistance, then he will descend the yellow line. If, on the contrary, the driver does not like interfaces that interact too much with him, he will set the line high enough. In the latter case of adjustment, we can for example place the yellow line at 80%, which makes it possible to be alerted just before the red threshold of 90% and thus avoid exceeding it.
Behavioral coaching and profiling
At the end of the trip, our Safety Coach offers statistics returned to the driver, which is complementary to proactive coaching. These data make it possible to comprehensively assess the behavior and the nature of driving: it is the driver’s signature.
Our technology calculates a detailed signature that we restore in a summary, partial and synthetic way, with a « gamming » aspect.
For example: (case of a real test run in Normandy).
An overall safe driving score.
We developed this rating with our panel of 3,500 testers: it varies analogicaly with the level of caution. We have tried a lot of settings, and have opted for a rating that always gives the impression of being rather high, even if we only win « cups » (bronze – silver – gold) from 92% (out of 100%).
Safety Coach and gamification
In order for the driver to want to play, he must be able to “win”. We have added a “facebook” button to our demo app: a sentence is automatically published on the user’s facebook (if they have accepted the connection), and only when the driver has won a cup: “I am a safe driver, today I won the bronze cup on a trip with the SafetyNex road safety tool ”. This part is there to suggest “gamification” or even incentive ideas to OEMs. An interesting incentive is the voucher at the car manufacturer brand’s store.
Safety Coach trip data
Also on our demo app (which is made to give ideas), we also offer the following data:
We are the only ones to offer this data:
- the safe driving score is the one displayed on the « cup » page
- anticipation is the driver’s ability to generate red alerts as little as possible
- the self-confidence expressed on a trip depends on the most frequent level of risk: if the driver regulates his risk around 60% he has expressed greater self-confidence than that of a driver who has regulated his risk around 50 %.
- driving skills: which represents the ability of a driver to reproduce his risk-taking identically throughout the trip. It is then easy to distinguish a pilot from a novice driver.
Tips for improving
Then, we propose a view of the risk on the duration of the trip:
The driver can visualize when he exceeded the yellow /orange threshold, and also threshold for red alerts, and what the difficulties were in front of the vehicle.
Finally, we give the driver a detailed coaching sentence:
This coaching phrase is “scalable” depending on the number of different variables that we want to be able to generate and the desired level of detail:
- the safe driving score
- the anticipation score
- the self-confidence
- the driving skills score
- the types of infrastructure that gave rise to an alert,
The processing of the above variables makes it possible to generate approximately 1.000 different sentences, and if we decide to generate 3 equivalent sentences (different formatting but same coaching) per case, we then have 3.000 sentences: this gives the feeling to the driver that vehicle is “talking to him”. The “robot” effect is completely erased. We noted during our panel test that this is very important for the experience of the user who does not like to find the same advice.
by L’Automobile & L’Entreprise
|“Président de Nexyad, qui a développé une IA cognitive globale, Gérard Yahiaoui nous présente le projet Oasis qui réunit aussi Yogoko, spécialiste des télécoms appliqués à la sécurité, et Milla Group, qui développe des navettes autonomes de dernière génération. Vers la fin des accidents ?”
“Toujours aussi brillant et pédagogue, Gérard Yahiaoui nous explique les détails du projet Oasis qui s’est formé dans le cadre de Nextmove (ex Mov’eo) : « Il faut le souligner car Nextmove est pleinement dans son rôle en créant des échanges entre des entreprises aux expertises variées et en faisant naître des projets ».
CEO of Nexyad, which has developed a global cognitive AI, Gérard Yahiaoui presents the Oasis project, which also brings together Yogoko, a specialist in telecommunications applied to safety, and Milla Group, which develops the latest generation of autonomous shuttles. Towards the end of accidents?
“Always brilliant and pedagogue, Gérard Yahiaoui explains the details of Oasis project which was formed as part of Nextmove (ex Mov’eo):“This must be emphasized because Nextmove is fully in its role by creating exchanges between companies with varied expertise and by giving birth to projects.”
Source et suite de l’article
Trois PME françaises élaborent une voiture capable d’anticiper les risques de la route.Trois PME françaises spécialisées dans l’automobile du future et l’intelligence artificielle cherchent à développer un programme capable d’analyser la qualité des routes, les conditions de trafic et les comportements des conducteurs afin d’anticiper la conduite à adopter et prévenir ainsi les accidents de la route.
Three French SMEs are developing a car capable of anticipating the risks on roads.Three French SMEs specializing in the automobile of the future and Artificial Intelligence are seeking to develop a program capable of analyzing the quality of roads, traffic conditions and driver behavior in order to anticipate the behavior to be adopted and thus prevent road accidents.
Données des véhicules : Nexyad salue les ordonnances gouvernementalesLa prise de position du ministère des Transports et de la Transition écologique a été reçue favorablement par Nexyad. Gérard Yahiaoui, CEO de Nexyad, parle de la fin d’une omerta sur les données, mais appelle à plus de précisions. C’est un petit pas pour le véhicule autonome, mais c’est un grand pas pour le législateur. “Le texte ne va pas assez loin, mais il va dans le bon sens“, a commenté Gérard Yahiaoui, le fondateur et CEO de Nexyad. Interrogé sur la portée des ordonnances annoncées la semaine passée par le ministère des Transports et de la Transition écologique qui comprend un volet sur le véhicule autonome, celui qui assure par ailleurs la vice-présidence de NextMove (ex-Pôle Mov’eo) en Ile-de-France a salué la volonté de “casser l’omerta des datas” tout en précisant qu’il reste à définir un cadre clair. “Nous sommes encore limités dans l’accès aux données qui permettent de nourrir l’intelligence artificielle et ce, parfois, pour de bonnes raisons, a rappelé Gérard Yahiaoui. Mais nos concurrents américains et chinois vont vite et il faut trouver un équilibre entre l’éthique …”
Vehicle data: Nexyad welcomes government orders
The position taken by the Ministry of Transport and Ecological Transition was favorably received by Nexyad. Gérard Yahiaoui, CEO of Nexyad, talks about the end of a data omerta, but calls for more details. This is a small step for the autonomous vehicle, but it is a big step for the legislator. “The text does not go far enough, but it goes in the right direction,” commented Gérard Yahiaoui, founder and CEO of Nexyad. Asked about the scope of the ordinances announced last week by the Ministry of Transport and Ecological Transition which includes a component on the autonomous vehicle, the one which also ensures the vice-presidency of French research cluster NextMove (ex-Pôle Mov’eo) in Ile -de-France welcomed the desire to “break the omerta of data” while specifying that a clear framework remains to be defined. “We are still limited in access to the data that feeds artificial intelligence, sometimes for good reasons,” said Gérard Yahiaoui. But our American and Chinese competitors are going fast and we have to find a balance between ethics … “
Lire la suite sur le JournalAuto
|Gérard Yahiaoui, CEO de Nexyad et VP de NextMove Ile de France|
Three French SMEs specializing in the automobile of the future and artificial intelligence are seeking to develop a program capable of analyzing the quality of roads, traffic conditions and driver behavior in order to anticipate the behavior to be adopted and thus prevent accidents. road accidents.
Read more on foodnewstoday24.com