AI Drives Efficiency for Automotive Management

Published: June 17, 2024
  • RFID, RTLS and IoT technology can feed a treasure trove of data to AI systems that manage finished vehicles.
  • This is the fourth in an ongoing series exploring how AI is impacting the RFID, IoT and different end uses across a number of industries.

Automative yards are notoriously complex. Even when things go as planned, managing finished vehicles at a manufacturing site, distribution yard or dealership, is time consuming and often inefficient. And that’s when things don’t go wrong.

In the real world, sometimes transport trucks don’t show up, cars don’t start, or a passing storm shuts down operations.

AI can make a company more agile and able to quickly respond to these events, but it needs data to understand what is taking place.

RFID, RTLS and IoT technology can feed a treasure trove of data to AI systems that manage finished vehicles. AI puts that data to use to better manage each vehicle, ensure it isn’t delayed —and in the long term—even plan around the vehicle’s entire processing without needing human oversight.

What a Vehicle’s Location Means

Consider a yard of finished vehicles awaiting transport to dealerships. If an AI system understands where each vehicle is located, and what that location means in its process, software can begin to forecast the next event for that car or truck and send alerts as parameters aren’t met.

Most companies are not yet using IoT data with AI for this purpose. But the value is going to make it more appealing, especially as the cost of the technology comes down, said Adrian Jennings, Cognosos’ chief product officer for AI and automotive.

Traditionally people have thought about real-time technology systems as a way to answer: “where’s my stuff?” Jennings pointed out. Companies with high value assets and inventory, including those who manufacture or own and operate vehicles, don’t want to lose track of them. So their goal has been: “if I lose this thing how can I find it quickly?” They attach an RTLS tag (transmitting via one of a variety of frequencies) and look up the location in their system software.

Inferring a Vehicle’s Status

The real value of RTLS comes with AI, to understand not just a car’s location but its place in a process.

“AI can accomplish process flow analysis and process flow optimization—that’s what the big data AI is great at,” Jennings said. “With AI, a company can ask questions such as not only ‘where is a car or van in my yard,’ but ‘where is it in my process?’”

One application is understanding—based on where a vehicle is parked—whether it is queued up for shipping, or in the next lane over,  awaiting finishing or specialized components.

“There could be two adjacent lanes, one row of cars that are all ready to ship and a row of cars right beside them that have to have some work done,” Jennings said.

Reducing Excess Driving

Companies with a fleet of vehicles to manage need to understand their processes and one commonly adopted solution with AI is tracking cars in in the outbound logistics, said Jennings.

Traditionally, such processing can be time-consuming as vehicles are moved from one part of the yard to another and back again. The cars might be driven to a processing center on site, go off site to a third-party accessories company to install specialty tires or spoiler, or are returned to a storage area to await the next step.

“Sometimes cars get moved more than they need to—into the processing center and back out and then again, from this part of the yard to this part of the yard and back again and then to the rail head and then back again,” Jennings said.

Each movement of a car comes with the cost of the labor for driving it, the gas, battery-use and risk of damage. For these reasons, Jennings pointed out, “you want to minimize the number of moves,” and that means having processes in place that anticipate each next move for one vehicle and for the other ones around it.

For AI, the low hanging fruit for vehicle management is to identify when and where a vehicle’s next step needs to be taken, and moving the vehicle at that time.

Saving Staging and Buffer Space

By reducing the shunting of cars around unnecessarily, an entire yard could be configured in a more efficient way that uses less space as well, Jennings said. For instance, cars that are scheduled for pick up are often located in general storage and brought to the truck loading buffer area.

The process of locating vehicles and bringing them to buffer can be time consuming as well as requiring additional dedicated space.

With AI, companies can leave the cars in their general storage space, as the car can be easily located and delivered to the truck at the time when loading is needed. This would eliminate the need for a buffer space.

Making Better Decisions in Advance

“Truthfully we’re really scratching the surface of what AI can do,” said Jennings. Most companies are using their RTLS tools for visibility that enables better human decisions on the ground. The next step is leveraging the system to make recommendations for optimized next steps. If something goes wrong, AI can be designed to find the next best process to correct the problem.

“In the long term, what you really want to do is feed a goal into an AI and have it take over from there,” said Jennings. Such a goal could be simply determining how 800 finished cars in a yard can be routed to the railyard over the course of a specific time or day.

AI can calculate the best schedule for movement, assign tasks to drivers in a staggered way, and therefore ensure that there are no traffic jams in the process.

“That would be like the dream scenario—where you could be feeding goals in the top and then sitting back and letting it just happen.”

EV Cars Increase Complexity

Dispatch planning has become more complex with the transition to electric vehicles (EV) cars as well, and AI can help manage the assignment of transportation accordingly.

EV’s are heavier than standard cars— due to their batteries—so selecting the right truck to pick them up, and calculating how many vehicles can be transported at a time, becomes more complicated.

An AI excels at optimization, to ensure trucks or railcars are filled to capacity, and that the EV cars are transported properly. “The ability to react to problems becomes the game-changer that keeps the process flowing at 95 percent when 100 percent isn’t achievable,” Jennings said.

Leveraging AI from Bottom Up

Additionally, AI can be used to help ensure the RTLS data is as accurate as possible, without spending a fortune on specialized sensors, something Cognosos specializes in.

Location accuracy of an RTLS system can be a challenge in environments that are filled with barriers and reflections like large metal containers and vehicles. Cognosos has developed deep learning algorithms that can understand and compensate for the errors caused by the trailers, said Jennings. “We can get very good tracking accuracy from very cheap sensors.”

In this way, Cognosos takes a two-sided approach to AI with its vehicle management technology. It uses AI from the “top down” on a server to assess what a vehicle’s status and next action should be, but it also relies on AI from the “bottom up” that assesses the location data captured from a low-cost wireless sensor.

Affordable Solutions

In the past, location accuracy caused challenges that limited some technology deployments, Jennings argues.

“The problem with IoT was always—how do you actually get the edge data that’s worth anything, that’s affordable enough, to actually deploy on scale to be worthwhile?” said Jennings.

Cognosos’ approach, however, enabled location accuracy from the sensor level, with AI, to provide more affordable solutions to collecting and analyzing data about vehicles. Jennings sees it as AI at the bottom that drives the AI at the top.

Learn More: