By Dr Prajwal Sabnis, Co-founder, Orxa Energies
Whether it is to accomplish a task as small as typing out an email or to oversee complex robotics employed for handling hazardous materials, Artificial Intelligence (AI) and Machine Learning (ML) are fast gaining ground as technologies with boundless potential. Industries across the board are employing these systems to reduce the need for human intervention in order to improve efficiency and reduce errors.
These solutions have immense potential for the electric mobility sector as well, where challenges like range anxiety continue to be major choke points that are slowing down the uptake momentum of Electric Vehicles (EV) today.
What Makes EVs Tick
The range of an EV is largely dependent on the energy density of its battery, along with the efficiency of its complimenting motor. Riding style and road conditions also play a significant role here, but with today’s battery technology, a range of about 150 km on a single charge is commonplace in the industry.
Batteries within the EVs are managed by a Battery Monitoring System (BMS) which is responsible for accurately monitoring vital parameters such as temperature, energy utilization, and charging.
Optimizing these systems requires the analysis of data in the real-time, and rapidly processing it into useful suggestions for the EV user to employ in order to use the vehicle to the full extent of the range possible from the battery.
AI and ML to the Rescue
Artificial Intelligence and Machine Learning algorithms analyze large amounts of data, spot patterns, and then process them into actionable suggestions.
That being said, the effectiveness of the output depends heavily on the quality of the data and the time it takes to process all of it. High-quality data can be collected by sensitive devices like sensors, accelerometers, and probes fitted on the vehicle itself, while external inputs can be collected through GPS, traffic camera grid, and the like.
Rapid developments in edge computing, in conjunction with efficient AI and ML algorithms, can drastically cut down the latency between data collection, computing, and generating an actionable output. AI and ML can then be employed to analyze riding patterns, road conditions, traffic density, and ambient temperature to suggest optimal riding inputs to the EV user, while simultaneously optimizing energy utilisation through the BMS. Motor efficiency can also be maximised by controlling its power output as per the riding conditions, and the regenerative braking can be employed to convert the kinetic energy of the brakes into electricity to extend battery range when the vehicle is braking or coasting.
This data analysis by AI and ML will also facilitate the display of a highly accurate range to the user computed as per the real time conditions, as opposed to the more rudimentary approximation that is prevalent today.
By constantly monitoring the vehicle’s location via GPS, ML and AI will enable the user to quickly find the closest recharging station when it is time to replenish the batteries in addition to reducing commuting time and distances by suggesting the quickest route to a destination. Furthermore, by hooking into the Advanced driver-assistance system (ADAS), AI and ML can be effectively used to pre-empt accidents, thereby reducing the probability of them happening.
All these facets enabled by AI and ML will drive efficiency, improve safety, and go a long way in reducing range anxiety amongst EV users and potential future buyers of these vehicles.
Aiding EV-related infrastructure
Just like with the vehicles themselves, AI and ML can help proliferate EV-associated infrastructure as well. By analysing data through suitable AI and ML algorithms, EV infrastructure providers can make informed decisions on where to set up charging stations and decide on the specifications of the hardware that is to be used. For example, in an area where the density of bikes is far more than cars, it would be more viable to set up more two-wheeler chargers than those for four-wheelers, leading to optimised usage of space and assets.
Once set up, AI and ML can maximise efficiency of these stations by analysing the charging patterns across the day and through the year. Depending on the charging load, AI and ML will enable the station to draw from the grid, or feed electricity back into it. This will make charging stations function like small electricity reservoirs that can help balance the electricity grid effectively by closely matching supply with demand.
As AI, ML, and the associated data acquisition and processing systems evolve, the potential benefits of them to the EV industry stand to increase exponentially. They will help EVs consistently deliver extensive ranges while keeping the user informed about the closest compatible charging stations and the like. As these technologies evolve, they will facilitate effective pre-emptive maintenance by analysing vital vehicle parameters like battery condition, powertrain health, along with the time interval and mileage covered between services.
Data processed by onboard ML algorithms can then be exported to external agencies like the government for future infrastructure development, eliminating bottlenecks for the smooth flow of traffic and identifying accident-prone areas and remedying them to be safer for road users.
OEMs too can use AI and ML to refine their products and make them better suited to fulfil the requirements of their customers and the markets they operate in. For CTOs and CXOs, the integration of AI and ML in the EV industry is not just a technological upgrade but a strategic imperative. It promises enhanced operational efficiency, innovative product development, and robust infrastructure growth, positioning the EV industry at the forefront of technological innovation and sustainable growth.