Get In: The Connected Vehicle Podcast From BlackBerry (Episode 4)
A major factor in the connected vehicle revolution is the rapidly accelerating switchover to electric cars. Battery power requires efficient management, including software that reacts to conditions both inside the car and in the wider environment. There is a huge opportunity here to leverage software to provide optimal battery range that reacts to location and driving style – with the right platform.
Meet Fabrizio Martini, co-founder & CEO of Electra Vehicles, a pioneer in the application of analytics and algorithms to optimize battery performance and improve vehicle range. Martini shared his vision for enlightened electrification of software-defined automobiles and more in the fourth episode of “Get In: The Connected Vehicle Podcast from BlackBerry,” our series that explores the possibilities created by -- and technologies behind -- the revolution in global transportation we are witnessing today.
The initial idea for Electra Vehicles came when Martini was working with NASA on a project to create an electric rover for exploring Venus. The challenges of making a vehicle work efficiently in such an inhospitable extraterrestrial environment seemed like it could pay dividends on our home planet, too. Martini, who also did stints at the U.S. Departments of Energy and Defense, says he “moved from Earth to space, back to Earth,” looking to apply what he had learned to more mundane modes of transportation.
The technology developed for the Venus explorer project had numerous applications. “NASA was looking to predict any potential failure of the energy storage, to make sure the vehicle was powered properly, as well as being able to talk to the station and to Earth,” says Martini. “So that's part of the connected vehicle aspect.” The NASA project was also interested in improving battery storage, as well as developing greater efficiency as the battery was used over time. These principles became a foundation for Electra’s core artificial intelligence technology (AI), “What we call EVE-Ai, our brain for batteries.”
However, electrification isn’t the only emerging opportunity for automakers. “We’re also going to start seeing more and more autonomous vehicles,” says Martini. “That is very linked to electrification because the two combined can bring very good efficiency, and that's what we need to utilize the battery with the greatest possible optimization. There's going to be a lot of connectivity between vehicles, so the vehicles will start to learn from each other.”
In anticipation of this convergence of autonomy with electrification, Electra has developed technology to help extend the range of electric vehicles. “EVE-Ai…can be deployed into any electric vehicle, any battery pack of any chemistry – from NMC, NCA utilized by Tesla, LFP, also innovative cell chemistries, like solid state batteries,” explains Martini. Electra’s team includes experts in machine learning and AI, data science, and software development. Together, they can develop an understanding of how each battery is going to behave in a given vehicle and with a specific driver, then adapt the control strategy accordingly.
“One of the features that we are deploying is called the EVE-Ai velocity recommendation. We coach the driver and give some advice on how fast to go to reach the destination,” Martini says.
“Another feature that we are bringing to market is called EVE-Ai dynamic state of charge,” continues Martini. “That gives the ability to push a button to receive an additional 50 miles, 100 miles, or more for a specific trip. We are seeing very good extended range, between 24 and 28 percent extra with a single charge, thanks to managing all the data that comes from the battery, from the vehicle, from the driver, and from the environment. There is great potential with the software layer on top of the batteries.”
The technology to power this push-button extension of battery life and driving range is applied so subtly and skillfully, Martini says, that the car’s occupants can’t tell the difference. “The experience to the driver is not perceptible. It's really a seamless adjustment of the control strategy. The EVE-Ai adaptive cell model and dynamic state of charge pushes the maximum-minimum limits of the batteries when it's safe to do so. There are certain times that when the battery is healthy, it's very relaxed, so it can push to the limit and give you that extra range.”
This advanced application of cutting-edge battery science also offers very positive implications for battery durability. “Most of the OEMs and Tier-1s are providing an eight-year battery warranty, sometimes 10,” says Martini. “But they would like 12, 15, 20 years. That's also important for the environment.”
“With our brain for batteries, we collect all the information around the batteries, the vehicles, the drivers, and the environment. We feed them into a neural network and recommend certain aspects, such as what should be the overnight charging curve. We have another solution called route optimization that helps to limit the wear of the batteries. A combination of these two features provides an increase in lifetime of about 30 percent.”
A fundamental enabler of this technology is the ability to collect and utilize all the data that the vehicle is producing from its proliferation of sensors, which is where Electra’s work with BlackBerry has become crucial. “We have started to collect more and more data,” says Martini. “BlackBerry’s innovative IVY platform acts as a middleware, helping to collect all this data from all these sensors. There is a big opportunity and we're going to see more and more applications like AI coming out – for methods of payment, for methods of recharging the batteries, methods of utilizing infotainment, and increasing safety.”
Electric motorsports provided an early proof point for Electra’s technology. “In 2015, one of our first case studies was on Formula E,” explains Martini. “We were trying to give extra range and extra power to the vehicles, and we saw great results. Formula E is improving year after year. The performance is going to surpass Formula 1. The question is when: Is it going to take 10 years, 15 years, 20 years? We don't know, but companies like BlackBerry with IVY and Electra with EVE-Ai could help performance improve.”
Another area where Electra’s EVE-Ai can pay dividends is in failure prevention. “Unfortunately, we still see large recalls,” says Martini. “We developed a software platform called EVE-Ai Predictive Analytics that helps to predict failure. It helps by telling which batteries can last one or two extra years to provide additional warranties to the driver or the fleet manager, versus the batteries that need to be recalled. You don't need to recall half a million batteries or 100,000 batteries. You can just recall a few thousand and take care of those, because those are the batteries that are having trouble.” This technology is greatly enhanced with a fleet of connected vehicles. “If you have a few thousand vehicles, they can start learning from each other.”
Martini sees two trends becoming the most influential over the coming years, thanks to connectivity. “Internet of Things is one big wave,” he says. “The other one is vehicle-to-X (V2X). That means vehicle to another vehicle (V2V); vehicle to infrastructure (V2I); or vehicle to grid (V2G).
“We started to see smart lights, for example, providing insights on whether the (traffic) light will be red or green, as well as buildings that will tell you if there are any available charging stations. Safety will be increased. No more shutdowns, recalls, thermal runaways, or dropping the range from 50 to 20 percent, just because it's cold outside.”
With such a bright future ahead, thanks to these connected vehicle innovations, Electra is very positive about its potential to grow. “We are in full hiring mode,” says Martini. “We are hiring talent from all over the world, all over the United States, and in particular, talent with expertise in machine learning, AI, and data science.” The connected vehicle is going to have some transformational improvements in store from Electra’s battery analytics technology, with a little bit of help from Venus.
You can subscribe to "Get In: The Connected Vehicle Podcast from BlackBerry" at the following media services:
Podcast Transcript Steve Kovsky: Fabrizio Martini: With the NASA team, we were working to develop a rover for exploration of Venus, an electric Rover. In that time, they needed help with the design of the battery packs, the analytics, in particular predictive analytics, as well as the control, in particular active control. I saw an opportunity to provide innovative solutions based on software technology, so I, with my co-founder and the team, started the business about five years ago from that NASA project. Steve Kovsky: Fabrizio Martini: That's where I started to work for the United States government. I worked for the Department of Energy, who were investing a lot in energy storage about 10-15 years ago and still do nowadays. Then I worked for the U.S. Department of Defense and finally at NASA. With the NASA team, we started to work on the project that I mentioned to you before. All pushing to electrification, application of batteries and energy storage. As I mentioned, about five years ago, we started to take some of that technology back to Earth, and that's why we set Electra at that time. Steve Kovsky: Fabrizio Martini: Steve Kovsky: Fabrizio Martini: The NASA team was looking to make sure that predictions for any potential failure of the energy storage, make sure to power the vehicle properly, as well as be able to talk to the station and to Earth. That's part of the connected vehicle aspect that we're going to talk about. Yes, a lot of concentration on the batteries and how to improve the energy storage over time. In fact, another aspect they were looking for is deploying an inactive control, a brain that could learn how the battery was deployed and utilized, and improve that over time. That’s what we develop at Electra and we call it EVE-Ai, our “brain” for batteries. Steve Kovsky: Fabrizio Martini: That is very linked to electric because the two combined can bring very good efficiency and that's what we need in order to utilize the battery at the most efficient that we can. I foresee more and more electric vehicles. This is an unstoppable wave. We are going to experience and drive and potentially purchase or lease an electric vehicle, as well as a combination with autonomous. We are seeing more and more self-driving vehicles with or without drivers that are going around. There is going to be the opportunity for people to rent or book or lease an electric vehicle or an autonomous electric vehicle as a service or owned. Then, I think also there's going to be a lot of connectivity between vehicles. The vehicles will start to learn from each other, will start to interact with each other, and that's the beauty of the technology coming up. I really can't wait to see that happen. Steve Kovsky: Fabrizio Martini: That helps to reach the destination, at the same time, but saving some energy from your batteries. We also have another feature that we are bringing to market. It's called the EVE-Ai dynamic set of charge. That gives you the ability to push a button and acknowledge that you're going to receive an additional 50 miles, 100 miles or more for a specific trip that you want to do. For a weekend that you want to go out with your kids or with your fiancé somewhere. That gives you the ability to extend the range of your electric vehicle for certain goals and certain extra mileage that you want. These are two of the features that we are deploying. We are seeing very good extended range between 24% to 28% extra range with a single charge. Thanks to managing all the data that come out from the battery, from the vehicle, from the driver and from the environment. Great potential with the software layer on top of the batteries. Steve Kovsky: Fabrizio Martini: Steve Kovsky: Fabrizio Martini: We feed them into a neural network, into this “brain” for the batteries and we recommend certain aspects such as the charging path. What should the overnight charge be? As an output, we recommend to the vehicle itself how to recharge the battery pack overnight. There is a strategy to improve the lifetime, as well as which path to take to reach your destination. We have another solution called route optimization that helps to limit the battery wear. A combination of these two features, overnight charging and route optimization, provide an increase of lifetime of about 30%. In certain case studies, with a combination of a secondary pack, we have seen up to twice the lifetime. Most of the time when we work on a single chemistry, we have seen 30% extended lifetime. That means 30% less cost of ownership as well. So, lifetime is very, very critical. Steve Kovsky: Fabrizio Martini: One way to do this is through edge computing, as well as cloud computing. That has been improving performance over time and establishing the opportunity. The relationship with BlackBerry that we have is due to the fact that BlackBerry has an innovative new platform called BlackBerry IVY™ that helps to act as a middleware. It helps collect all this data from sensors that we’ve been talking about to improve and create the new synthetic sensors that can help develop applications and protocols, like the ones we are talking about for the velocity recommendation, dynamic set of charges and so on. The way to collect data, compute some of those data for sensors and synthetic sensors and come up with value. Value to the driver, value to the OEM, to the Tier 1 and to the vehicle itself. I believe there is a big opportunity and we're going to see more and more applications similar to EVE-Ai coming out for method of payments, for method of recharging the batteries, method of utilizing infotainment and more, and increasing the safety as well. Steve Kovsky: Fabrizio Martini: There is a very big opportunity with that aspect. Lately, what we have seen is that Formula E started to ramp up. I'm glad to see more and more opening from the FIA that showcase and give a little bit more liberty to companies to explore different battery technologies, different control algorithms, different powertrain design and strategies. I believe there is a big opportunity there. Just an insight when we started Electra Vehicles in 2015, one of the first case studies that we've done was on Formula E. We were taking a couple of trucks of that championship, and we were taking the power profile of that. We were applying some of the knowledge and insight that we transfer from NASA on how do we manage the battery? How do we optimize the batteries? How would we give a little bit of extra range? We still see a couple of vehicles that are not able to complete all the laps, the 8, 9, 10 laps of Formula E. We were trying to give the extra range, give that extra power to the vehicles and we've seen great results. One of the first case studies for Electra was actually applied to Formula E. That's quite interesting. I think we're going to see more and more advancement there and I can't wait to see what they come up with. Steve Kovsky: Fabrizio Martini: Steve Kovsky: Fabrizio Martini: Most of the time, you don't need to recall half a million batteries or 100,000 batteries. You can just recall maybe a few hundred or a few thousand and take care of those, because those are the batteries that are having some trouble. Thanks to EVE-Ai Predictive Analytics, we can spot those batteries and predict failure. We have demonstrated this with a couple of our clients in the commercial vehicle industry, where we could predict a failure three to six months in advance. I think that's one of the challenges we are still seeing. Steve Kovsky: Fabrizio Martini: What we have seen is also the benefit aspect. Everybody that is trying an electric fleet starts cutting down the cost. That's something very important for the fleet manager. In addition, there is a connected mobility aspect. Whoever has a fleet of vehicles, can start to learn from each other. If you have a few thousand vehicles, your vehicles can start learning from each other, where are vehicles behaving better in a geographical area versus another? Maybe some drivers drive electric vehicles better than others, are less aggressive. That's why we came up with green leaf scoring cards and points to the better driver for EVs. I think there is a big opportunity here, particularly for the commercial fleet and connected mobility. Steve Kovsky: Fabrizio Martini: We started to see smart lights, for example, providing insight on whether the light will be red or green, as well as buildings that will tell you if there are any available charging stations and if so, tell the vehicle to go there and recharge at a certain temperature and current. That's something that we've seen more and more. Our approach at Electra is also taking some of the input and feeding them into our brain for batteries, in our EVE-Ai. That's why we call it a 360-degree solution. It's really 360 because we capture data where we can and one of those aspects is the vehicle-to-x, the vehicle to infrastructure, other vehicles and greet as well. Steve Kovsky: Fabrizio Martini: Steve Kovsky: Fabrizio Martini: Steve Kovsky: |