As we can see, the automotive industry is undergoing a major transformation with the electrification of vehicles, driving aids and the arrival of the first autonomous vehicles in major US and Chinese cities. But there is also a strong enthusiasm for the adoption of generative AI, with three-quarters of companies already experimenting with applications.
Indeed, the sector estimates that generative AI could improve R&D processes by 10 to 20%. Companies are looking for productivity gains and cost reductions, particularly for:
– software testing and validation
– acceleration of test and validation, marketing and quality.
However, organizational transformations are necessary to maximize this value.
In United States, the car sharing market is projected to grow significantly by 2025, according to analysts a fourfold increase in fleet size to about 427,000 cars and a fivefold rise in users, reaching 36 million. This growth is driven by government support, technology advancements, and service optimization.
Key technologies like artificial intelligence and machine learning will enhance customer personalization, allowing tailored rates and recommendations. They will also improve user experience through features such as customized vehicles and integrated route planning with other transport modes.
For businesses, these technologies aid customer retention by enabling targeted offers and demand forecasting. Proactive measures will address driving behaviors, implementing restrictions for aggressive drivers and ensuring vehicle maintenance. As competition intensifies, car sharing companies will focus on enhancing service quality and expanding their feature offerings.
After more than 15 years of work thru 12 government funded collaborative research programs with experts of road safety, police of the road, professional drivers and insurers of 19 countries, Nexyad has developed a Metric of Driving Prudence.
Using this Metric of driving Prudence, Nexyad offers SafetyNex an AI tool on board and real time which analyses driving prudence 20 times per second, comparing driving behavior (speed, accels) with road context ahead (road geometry and signs, obstacles, weather, visibility, etc.). When we detect lacks of prudence, we are able to alert drivers with few seconds of anticipation in order for them to slow down and then reduce accident situations.
« The tool, called Tutor CoPilot, demonstrates how AI could enhance, rather than replace, educators’ work.
Tutor CoPilot isn’t designed to actually teach the students math—instead, it offers tutors helpful advice on how to nudge students toward correct answers while encouraging deeper learning. »
A research study by National Student Support Accelerator highlights the potential of Generative AI in education, focusing on Tutor CoPilot, a Human-AI system that aids tutors. A study with 900 tutors and 1,800 K-12 students showed that using Tutor CoPilot increased student mastery in math by 4 percentage points, especially benefiting lower-rated tutors (9 percentage points). At just $20 per tutor annually, it promotes effective teaching strategies. Despite some limitations in suggestions, the findings suggest Tutor CoPilot can improve educational quality and accessibility for underserved communities.
First steps have been taken to, perhaps, eventually identify what is produced by a generative AI and distinguish it from the real production of a human. The latest comes from a major player Google.
Obviously, this can only come from the companies that develop these AIs. It would make no sense for a user who shares content to denounce themselves, especially if there is a gain at the end, whether financial, prestige or authenticity.
Is GenAI a tool like any other? Like the calculator or the electric motor ?
« Making GenAI Better for You: Experimentation for Custom Research Solutions »
Shane Storks, Graduate Student Research Assistant, Computer Science and Engineering, College of Engineering.
Michigan Institute for Data Science and AI Laboratory
India is a nuclear power, with the bomb. It is the 5th largest economy in the world ahead of the United Kingdom, the most populous country and a democracy with almost a billion voters. In the future, India could push global growth with such a huge and rapidly awakening market.
In such context, transportation is great challenge for acceleration and India intends to propose ADAS solutions to combat road insecurity. Indeed, 173,000 deaths in 2023, more than 400,000 injured on the roads.
Of course, there is a lot of work to do in this large country in terms of infrastructure. Roads are not proper everywhere and drivers sometimes need to be helped for greater safety. Motiv AI a company from Coimbatore is starting to deploy such assistance with a smartphone App powered with AI that works in real time to alert all vehicle drivers with anticipation when they make lacks of prudence.
Nexyad Dreamotor1 vehicle integrates Hybrid AI SafetyNex which performs on-board analysis of the Driving Behavior in relation with Road Context, in real time (20 times/s). SafetyNex allows to Anticipate singularities on road in order to assist drivers or pilot a predictive ACC.
The exchange of data between the vehicle and other entities (other vehicles, road infrastructure, vulnerable road users such as pedestrians or 2-wheelers and the cloud) is performed using YoGoKo’s unified communication & data management software platform (Y-SMART).
The demonstration here shows reaction of SafetyNex to an incoming message originated from some Hazard Warning info platform, using V2X communications.
The message informs that a motorcycle (virtual) is laying on ground after an accident. When Dreamotor1 is informed of the hidden accident at some known distance ahead, it provides the information to the driver (asking to slow down), or it reduces automatically speed in ACC mode.
V2X information is a possible input of SafetyNex (traffic, accident, road work, weather, etc.)
Other inputs are eHorizon (Map SD or HD); Diagnosis of detection sensors (obstacle and DMS)
This year, Paris Mondial de l’Auto has regained more manufacturers, more visitors than previous edition. People still seems to love cars, they make kids to seniors dreaming. We could feel electricity on the air because it is everywhere. A lot of new models from French brands on super design booths. Chinese cars almost occuped one of the four pavillons of Paris convention center, with more sedan’s than city cars, even exentric premiums.
Paris Mondial will return in october 2026.
After seeing AI companies attract more and more money (article here), there is an increasing financial investments by American tech giants – Microsoft, Google and Amazon – who are positioning themselves for the future by choosing nuclear energy.This move is particularly strategic because nuclear energy stands out as an energy source that does not produce CO2 emissions while providing a reliable and constant supply of electricity.As the computing requirements of the AIs they develop continue to increase, these companies are committing billions of dollars to meet their energy needs. Those who do not have access to sufficient energy resources could suffer in the competitive landscape of AI development.
Milan Kovac, VP Head of Engineering Optimus at Tesla shared a couple of news after 10/10 event held at Warner Bros. studios in California.
Here a summerize:
Tesla Bot recently held an event where it showcased the progress of its Optimus robots. About 20 active robots chatted with visitors for about four hours at the 10/10 event. The robots demonstrated impressive balance and mobility, and there was only one minor incident. Highlights included the unveiling of a new generation of hand and forearm that greatly improves dexterity and sense of touch.
Work continues, and marks advances in the robot’s autonomy: Optimus is now able to visually navigate indoors using its 2D cameras. It can avoid obstacles, navigate autonomously, carry payloads, and dock to recharge. Efforts to improve autopilot technology, used in both cars and robots, have contributed to these improvements.
Additionally, Optimus is being trained to interact with humans and respond to gestures and voice commands to give them snacks and drinks. There is certainly still a lot of work to be done, but there is optimism by Elon Musk’s teams about future developments.
For the second time, after 2022, PFA (Platform of French Automotive) organise its Summit during the Paris Mondial at Paris Convention Center, Porte de Versailles in Paris.
Concern and determination are the two impressions that were revealed by the various top leaders’ interventions (see list below)
The European automotive industry is facing a drop in sales on the old continent that is multifactorial:
– General economic slowdown
– Increase in energy prices
– Demographic decline of potential buyers
– Arrival of Chinese vehicles
– Cut-throat regulation that could worsen the situation
But the manufacturers, tier one suppliers and side players, back to the wall, want to be combative. They have the collective will to organize themselves on several items:
– Innovation (batteries, Software Defined Vehicle, ADAS, AI)
– Industrial collaboration at a European scale
– Alliances with Asian players
– The call to EU political bodies for support adapted to current and future realities of sector
Introduction by Luc Chatel President of PFA
Keynotes :
Luca de Meo, CEO Renault Group
Oliver Zipse, CEO BMW Group
Christophe Perillat, CEO Valeo
John Bozzela, President of the International Organization of Motor Vehicle Manufacturers (OICA), president and CEO Alliance for Automotive Innovation (USA)
Carlos Tavares, CEO Stellantis
Interview :
Patrick Pouyanné, CEO TotalEnergies
Christel Heydemann, CEO Orange Group
Closing : Antoine Armand, Minister of Economics, Finance and Industry of France
Use-Case that described a common situation leading to many accidents:
Pedestrian crossing partly masked by a bus, a pedestrian surrounds from behind the bus.
As the vehicle is approaching the pedestrian crossing, it slows down slightly (a prudence function is activated at the crossing, which can be disabled if you believe your customers prefer that, similar to other Prudence Functions in the tool). We could also incorporate contextual information to determine if the area is urban or peri-urban. The vehicle detects the parked bus, geolocates it, and recognizes that it is partly masking the view of the pedestrian crossing, prompting a reduction in speed even if no pedestrian is currently visible. When the pedestrian does appear, the vehicle can stop at a deceleration of 0.5 g, avoiding the activation of the Automatic Emergency Braking (AEB) system.
Please note that we cannot calculate the potential collision with a pedestrian because potential pedestrian is masked by the bus.
The pilot system is a PID (you can utilize your existing pilot).
Understand Vehicle HMI:
See video : simulation with AURELION (dSpace)
The vehicle is driven by SafetyNex, on a real pathway around Nexyad office (digital twin).
« Who Is That You Are Chatting With? Oh, Just ChatGPT.
The artificial intelligence chatbot’s Advanced Voice Mode feature has delighted some users and weirded out others.
When you think of what a voice programmed by artificial intelligence would sound like, you might picture something robotic and stilted, with a staccato cadence incapable of capturing the inflections, speed and emotion required to sound even somewhat human. But this is 2024, and the robots have gotten a serious upgrade. Now they can imitate voices, accents and intonation to an almost creepy degree — for better or worse. »
For better or worse ? The both of course.
This could probably be nice for a typical user looking for information, ordering a meal, or making a doctor’s appointment, and even support for a single person who no longer communicates with others. But it also risks generalizing deep fake vocals.
When cavemen banged stones on top of twigs to warm themselves with fire, it was a great innovation, a step forward for the human kind. And then they discovered other less positive uses…
Nexyad’s innovation, BYOD* SafetyNex, in its informative driving assistance mode for fleets and/or insurance companies, consists of warning drivers on board when they are not being prudent behind the wheel in real time. SafetyNex can then launch simple messages by anticipation such as: « slow down » or even contextualized « slow down dangerous bend », etc.
Our tool enhanced with ChatGPT-type technologies could then launch alerts such as « Honey, please slow down » in your wife’s voice, or « Daddy, go slower » in your child’s voice.
In this combination, our 100% positive innovation will make the vocal imitation for the better way. *Bring Your Own Device
It takes time in science to estimate the scope of important work.Finally, John Hopfield born in 1933 and Goeffrey Hinton in 1947, have just jointly obtained the 2024 Nobel Prize in Physics for their work carried out since the 80’s on machine learning.
Goeffrey Hinton is known for his applications:
Backpropagation in machine learning is a gradient estimation method commonly used for training neural networks to compute the network parameter updates.
Boltzmann machine, a statistical physics technique applied in the context of cognitive science.
Deep learning (considered as one of the godfathers) is a subset of machine learning methods based on neural networks with representation learning.
Capsule neural network (CapsNet) is a machine learning system that is a type of artificial neural network (ANN) that can be used to better model hierarchical relationships. The approach is an attempt to more closely mimic biological neural organization.
John Hopfield developed :
Hopfield network (or associative memory) is a form of recurrent neural network, or a spin glass system, that can serve as a content-addressable memory.
Modern Hopfield networks(also known as Dense Associative Memories) are generalizations of the classical Hopfield networks that break the linear scaling relationship between the number of input features and the number of stored memories.
Andrej Karpathy joins Sarah and Elad in this week of No Priors. Andrej, who was a founding team member of OpenAI and the former Tesla Autopilot leader, needs no introduction. In this episode, Andrej discusses the evolution of self-driving cars, comparing Tesla’s and Waymo’s approaches, and the technical challenges ahead. They also cover Tesla’s Optimus humanoid robot, the bottlenecks of AI development today, and how AI capabilities could be further integrated with human cognition. Andrej shares more about his new mission Eureka Labs and his insights into AI-driven education and what young people should study to prepare for the reality ahead.
Complexity Reduction Technology, Gain of Time and Money
SafetyNex by Nexyad is a technology that reduces with 100 factor (at least) complexity of development, test and validation for Highly Automated vehicle or HAV.
We do not use Petaoctets of recorded data for treatment in computers farms large as malls. Nexyad worked on a new paradigm based on experience of prudent professional drivers, experts on road safety, and info from insurers, building a metric of prudent driving. Ordinary PC runs our technology using 5% of CPU.
SafetyNex doesn’t identify scenarios, but works on separated elements or events that make up scenarios. We associate a PRUDENCE FUNCTION to each detected element/event in parallel. Prudence functions are dynamic as elements/events are. We fusion them in real time with knowlegde-based, fuzzy logic and possibility theory AI. Doing so, SafetyNex replicates the way humans manage prudence/risk.
Use-Cases Simulation Video of Safetynex
The road, shown in video below, really exists in France near Nexyad office. AI SafetyNex is interfaced with RTMaps of Intempora dSPACE Company in simulation sensors tool AURELION by dSPACE. We just changed trees and houses, but alignment of the road is the same than real world. The legal speed limit is 80 km/h or 50 mph all the way.
Nexyad SafetyNex controls longitudinal and lateral of vehicle. HMI of autonomous vehicle on video shows legal speed sign, current speed on the left, square of prudence (green, yellow or red) with next speed to apply at tapering given distance, and the reason by road sign. Just below, the prudence target accepted: here we have set the threshold of driving prudence at 50% (range from 0 to 100). This corresponds to an occasional driver, not very experienced.
In the following video (eagle & cockpit views), in two use-cases, elements/events are:
a) use-case Winding Curves (prudence function)
b) use-case Curve (prudence function 1) with No Visibility (prudence function 2) In the second use case (curve with no visibility) two prudence functions are merged*.
So AI SafetyNex manages autonomous driving, reading and analysing a map thru eHorizon and detects winding curves ahead in use-case a, deciding prudent speed at n meters.
Then use-case b, curve (prudence function 1) with no visibility (prudence function 2) is detected both on map (Here Technologies) and by Aurelion sensors (camera). SafetyNex reduces speed anticipating possible object hidden ahead. In this example there a very slow truck on the road. Once truck detected, SafetyNex turns red and decides to brake. Notice that collision was avoided because of the previous slowing down of prudence functions 1 and 2. With the legal speed of 80 km/h or 50 mph, car would have ended up in the truck or in the scenery.
Look at this simple seemingly use-case and think how it can be imagined and recorded in this particular context (this type of curve there) by classical method of scenarios…
*if you want to challenge SafetyNex, and see several prudence functions merged to make a complex situation in urban area for example, contact us.
Money for AI, whereas Open AI creator of ChatGPT raises new $ 6,6 billion, Startup Poolside that aims to create AI capable of “human-level intelligence” in software developmentpockets $ 500 million at B serie, and HPC-AI Tech specializing in AI software infrastructure and video generation gets a first round of $ 50 M.
We have developed an AI technology for vehicle fleets (trucks, cars, two-wheelers) that measures driver prudence in real time. Where our competitors measure possible indirect consequences of poor driving, such as harsh braking, we offer a real prudence metric. This metric was developed through 12 research programs co-financed by the French government, and is currently being deployed in India. The company MOTIV AI from Coimbatore has in fact developed a smartphone App based on our technology SafetyNex.
Thanks to this App, fleets reduce the number of accidents and therefore reduce their operating costs. They can also train their drivers to be more careful.
Stellantis Chief Software Officer Yves Bonnefont, interviewed by S&P Global Mobility about Software-Defined-Vehicle technology.
Challenges, Issues and Strategy.
– Key factors driving Stellantis to develop SDVs.
– SDV architecture transformations and their potential impact on vehicle operations
– The division of software development between Stellantis and suppliers?
– Developing more software in-house that supports more of the platform, API and parent stack, or even deepening the software stack
– Organizational challenges related to SDV development
– Situations where hardware upgrades can be performed in the field
– Development of SDVs for the entire product portfolio or mainly for BEV platforms or the premium segment
– Security precautions, especially for SDVs that have deeper cloud connections
We use cookies to understand how you use our site and to improve your experience. By continuing to use our site, you accept our use of cookies. AcceptPrivacy & Cookies Policy Read More
Privacy & Cookies Policy
Privacy Overview
This website uses cookies to improve your experience while you navigate through the website. Out of these, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may affect your browsing experience.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.