Demand for food is surging, with projections indicating that crop production must nearly double by 2050 to sustain the global population. Chemical crop...

The post AI is About to Massively Accelerate Innovation in Crop Inputs appeared first on Cleantech Group.

Car design is an iterative and proprietary process. Carmakers can spend several years on the design phase for a car, tweaking 3D forms in simulations before building out the most promising designs for physical testing. The details and specs of these tests, including the aerodynamics of a given car design, are typically not made public. Significant advances in performance, such as in fuel efficiency or electric vehicle range, can therefore be slow and siloed from company to company.

MIT engineers say that the search for better car designs can speed up exponentially with the use of generative artificial intelligence tools that can plow through huge amounts of data in seconds and find connections to generate a novel design. While such AI tools exist, the data they would need to learn from have not been available, at least in any sort of accessible, centralized form.

But now, the engineers have made just such a dataset available to the public for the first time. Dubbed DrivAerNet++, the dataset encompasses more than 8,000 car designs, which the engineers generated based on the most common types of cars in the world today. Each design is represented in 3D form and includes information on the car’s aerodynamics — the way air would flow around a given design, based on simulations of fluid dynamics that the group carried out for each design.

Each of the dataset’s 8,000 designs is available in several representations, such as mesh, point cloud, or a simple list of the design’s parameters and dimensions. As such, the dataset can be used by different AI models that are tuned to process data in a particular modality.

DrivAerNet++ is the largest open-source dataset for car aerodynamics that has been developed to date. The engineers envision it being used as an extensive library of realistic car designs, with detailed aerodynamics data that can be used to quickly train any AI model. These models can then just as quickly generate novel designs that could potentially lead to more fuel-efficient cars and electric vehicles with longer range, in a fraction of the time that it takes the automotive industry today.

“This dataset lays the foundation for the next generation of AI applications in engineering, promoting efficient design processes, cutting R&D costs, and driving advancements toward a more sustainable automotive future,” says Mohamed Elrefaie, a mechanical engineering graduate student at MIT.

Elrefaie and his colleagues will present a paper detailing the new dataset, and AI methods that could be applied to it, at the NeurIPS conference in December. His co-authors are Faez Ahmed, assistant professor of mechanical engineering at MIT, along with Angela Dai, associate professor of computer science at the Technical University of Munich, and Florin Marar of BETA CAE Systems.

Filling the data gap

Ahmed leads the Design Computation and Digital Engineering Lab (DeCoDE) at MIT, where his group explores ways in which AI and machine-learning tools can be used to enhance the design of complex engineering systems and products, including car technology.

“Often when designing a car, the forward process is so expensive that manufacturers can only tweak a car a little bit from one version to the next,” Ahmed says. “But if you have larger datasets where you know the performance of each design, now you can train machine-learning models to iterate fast so you are more likely to get a better design.”

And speed, particularly for advancing car technology, is particularly pressing now.

“This is the best time for accelerating car innovations, as automobiles are one of the largest polluters in the world, and the faster we can shave off that contribution, the more we can help the climate,” Elrefaie says.

In looking at the process of new car design, the researchers found that, while there are AI models that could crank through many car designs to generate optimal designs, the car data that is actually available is limited. Some researchers had previously assembled small datasets of simulated car designs, while car manufacturers rarely release the specs of the actual designs they explore, test, and ultimately manufacture.

The team sought to fill the data gap, particularly with respect to a car’s aerodynamics, which plays a key role in setting the range of an electric vehicle, and the fuel efficiency of an internal combustion engine. The challenge, they realized, was in assembling a dataset of thousands of car designs, each of which is physically accurate in their function and form, without the benefit of physically testing and measuring their performance.

To build a dataset of car designs with physically accurate representations of their aerodynamics, the researchers started with several baseline 3D models that were provided by Audi and BMW in 2014. These models represent three major categories of passenger cars: fastback (sedans with a sloped back end), notchback (sedans or coupes with a slight dip in their rear profile) and estateback (such as station wagons with more blunt, flat backs). The baseline models are thought to bridge the gap between simple designs and more complicated proprietary designs, and have been used by other groups as a starting point for exploring new car designs.

Library of cars

In their new study, the team applied a morphing operation to each of the baseline car models. This operation systematically made a slight change to each of 26 parameters in a given car design, such as its length, underbody features, windshield slope, and wheel tread, which it then labeled as a distinct car design, which was then added to the growing dataset. Meanwhile, the team ran an optimization algorithm to ensure that each new design was indeed distinct, and not a copy of an already-generated design. They then translated each 3D design into different modalities, such that a given design can be represented as a mesh, a point cloud, or a list of dimensions and specs.

The researchers also ran complex, computational fluid dynamics simulations to calculate how air would flow around each generated car design. In the end, this effort produced more than 8,000 distinct, physically accurate 3D car forms, encompassing the most common types of passenger cars on the road today.

To produce this comprehensive dataset, the researchers spent over 3 million CPU hours using the MIT SuperCloud, and generated 39 terabytes of data. (For comparison, it’s estimated that the entire printed collection of the Library of Congress would amount to about 10 terabytes of data.)

The engineers say that researchers can now use the dataset to train a particular AI model. For instance, an AI model could be trained on a part of the dataset to learn car configurations that have certain desirable aerodynamics. Within seconds, the model could then generate a new car design with optimized aerodynamics, based on what it has learned from the dataset’s thousands of physically accurate designs.

The researchers say the dataset could also be used for the inverse goal. For instance, after training an AI model on the dataset, designers could feed the model a specific car design and have it quickly estimate the design’s aerodynamics, which can then be used to compute the car’s potential fuel efficiency or electric range — all without carrying out expensive building and testing of a physical car.

“What this dataset allows you to do is train generative AI models to do things in seconds rather than hours,” Ahmed says. “These models can help lower fuel consumption for internal combustion vehicles and increase the range of electric cars — ultimately paving the way for more sustainable, environmentally friendly vehicles.”

“The dataset is very comprehensive and consists of a diverse set of modalities that are valuable to understand both styling and performance,” says Yanxia Zhang, a senior machine learning research scientist at Toyota Research Institute, who was not involved in the study.

This work was supported, in part, by the German Academic Exchange Service and the Department of Mechanical Engineering at MIT.

© Credit: Courtesy of Mohamed Elrefaie

The New York Times referred to the major IT outage in July involving Microsoft and CrowdStrike as the “glitch felt around the world.” In today’s digital age amid an increased presence of artificial intelligence tools, it’s no surprise that keeping sensitive data safe is a rising concern for the student transportation industry.

During his STN EXPO West keynote presentation in July, Keith Krueger, the chief executive officer for the Consortium for School Networking (CoSN), discussed the results of an annual survey of school IT leaders that indicated a shift in the top five technology priorities. The number one challenge for the past three years was cybersecurity. Data Privacy & Security, which had been sitting at No. 5 in 2022, moved up to No. 2. Network Infrastructure was third followed by the up-and-coming challenges of Determining AI Strategy and IT Crisis Preparedness.

Transportation departments are handling large amounts of data, including but not limited to onboard video camera footage, student ridership verification, telematics, and transportation employee information.

With these large amounts of data, it’s also important that school districts and vehicle contractors are equipped to effectively use and analyze the data, which could increasingly incorporate the application of AI.

Different facets of data and information security were discussed at the 2024 STN EXPO West conference in Reno, Nevada, in July. During these sessions, speakers and vendors discussed how increased technology offerings also require due diligence in protecting the data that is run through and stored in a given solution.

In one of the STN EXPO West sessions, representatives from Geotab and Tyler Technologies led a discussion titled “What Do I Do with All of this Data? Using Artificial Intelligence and Business Intelligence tools in Student Transportation.” Craig Berndt, the segment manager for student transportation at Geotab, noted that he is expecting AI to be a rising trend.

“Machine learning is like teaching your dog to fetch, except your dog is a computer and fetching is recognizing patterns in data,” he explained. Some of the applications using AI to track data discussed in the sessions included risk management, predictive maintenance, driver training, tracking student attendance, and continuous learning that can assist in effective routing planning.

Berndt noted that historically there has been much conjecture surrounding AI, and a lot of that is hyperbolic. Geotab displayed its new AI assistant software Geotab Ace at the STN EXPO West Trade Show. Berndt added that Geotab protects transportation data by keeping it on a private, secure server. He explained it’s important to know how your data tools work and exactly where the data is landing.

“No one here would put your student data into ChatGPT. Our goal with generative AI is to get away from the staff having to analyze reporting. Would you like to be told what trends are from a reliable source or have to go through the data yourself?” he noted.

Berndt said that it’s important to stay on top of trends in AI, data security and analysis, commenting that “Artificial intelligence isn’t going to take away your job. People who know how to use artificial intelligence are going to take your job.”

Protecting sensitive student data was the topic of a panel discussion moderated by Rick Hays, deputy chief information officer at the Nevada Department of Transportation. Hays holds a doctorate in cybersecurity, served in the U.S. Air Force, and worked for the Cybersecurity and Infrastructure Agency, an arm of Homeland Security. He has extensively worked on military and government levels to further cybersecurity safety practices.

Panelist Jennifer Vobis, who has since retired as executive director of transportation for Clark County School District in Nevada, spoke about a 2020 security breach that affected 40,000 district employees. It wasn’t until three years later that the district discovered information had been sold on the dark web. Vobis said that while her department assumed IT had the data security covered, it’s important to fully understand how those imperfect safeguards affect transportation operations.

Hays noted that many ransomware attacks begin with an email, an easy-to-overlook threat. His advice was to take a moment to analyze the message and sender, and “think before you click.”

“Balance the drive to get tasks done with making sure we know what it is we’re doing,” added panelist Lam Nguyen-Bull, a consultant at Edulog and an attorney, explaining that it’s everyday behavior that creates the most risk.

She continued that understanding and managing data flow and security starts with understanding that “data is just information,” whether physical or digital. Just as Berndt noted, Hays also emphasized the importance of knowing exactly where data is at all times. When it’s being used, when it’s being stored and when it’s in transit. Encryption must be present at all these levels, he explained.

Nguyen-Bull continued that data in storage is the easiest stage to protect it. When data is in transit across the web, it is generally protected by a Secure Sockets Layer (SSL) or Hypertext Transfer Protocol Secure (HTTPS). When it’s being used it is protected by a firewall in a closed environment.

“What makes it vulnerable is when it’s between stages,” she said. “When it’s not being managed by a system.”

Nguyen-Bull used the example of a parent portal app, which she referred to as “a perfectly safe product if you use it right.” Ensuring that only the relevant parties can view data or a particular school bus location, or that a tablet onboard the bus is locked is the responsibility of the owners of the data. “Know what your responsibilities
are regarding the data you handle, you need to know the policies,” said Hays.

The human element of safely managing and effectively protecting data is a team effort, said Vobis, but it may be a teaching moment if all the staff is not up to date on technological education and cybersecurity training. Even though some of these practices may be considered common sense, the panel stressed the importance of covering all your bases and making sure each member of the team understands the implications of data breaches.

When things go wrong, and Nguyen-Bull noted that they will, it’s crucial to have an action plan in place to not only get the issue under control but to understand what happened and how it can be prevented in the future.

During a security audit situation, like one a “white hat” firm performed on Edulog last year, “We don’t usually like to answer questions, but understand we’re not being attacked. [Auditors] are just trying to understand,” she explained. “Be collaborative, learn from other people’s experiences. Despite best efforts to lock things down, there is always a high risk.”

We always think it’ll never happen to us,” said Vobis. Even after the situation at Clark County was resolved, she said there was an impact on how information was shared. Vobis cited an example of improper information sharing via Google Suites, where security privacy settings weren’t on. Nguyen-Bull referenced receiving an email with an attached unencrypted spreadsheet containing detailed data on student riders.

“Practice doesn’t make perfect, but practice does make it better,” said Nguyen-Bull, recommending that districts run tabletop exercises to prepare for when the “unthinkable does happen.”

All the panelists advised that student transporters take time to find out their organization’s cyber policies and security protocols.

Hays spoke to the widespread variety and type of ransomware and cyberattacks, noting that they can happen to very small and extremely large organizations, alike. He advised that transportation departments should have software in place to scan incoming files for possible attacks and that transportation should coordinate with the district to ensure security protocols for transferring or receiving data is being upheld throughout all operations.

Nguyen-Bull noted that even though it may seem like data is spread out between multiple people or databases, it can be easy for that information to get centralized somewhere within the district. She continued that predictive computation could use any amount or type of data to create complete pictures.

“Data is permanent, in all forms,” said Hays. “It can come back to bite you, no matter what stage it’s in.”

Both he and Nguyen-Bull advised being cautious with “new and improved AI” technology that is being created to meet the demand of ever-increasing data. Hays referenced the addage “Trust but verify,” which he said is applicable to all of us, in our personal and professional lives.

In a continually evolving digital landscape, Nguyen-Bull said that while she does work for a software company, she makes sure to prioritize people with face-to-face and voice interactions.

“Don’t reduce everything to digital.”

Editor’s Note: As reprinted in the November 2024 issue of School Transportation News.

---

Related: What Do You Really Need from Technology?
Related: (STN Podcast E218) Onsite at STN EXPO Reno 2024: Coming Together for Safety, Technology & Clean Energy
Related: STN EXPO INDY AI Session Advises Attendees to be Curious but Cautious
Related: Technology and Communication: Crucial for Bus Maintenance and Safety

The post Managing Transportation Data and Keeping It Safe appeared first on School Transportation News.