Each year, billions of tires are produced globally. Mounting these tires and performing periodic tire rotations requires precisely aligning lug nuts to wheel studs, the threaded fasteners that provide wheel-mounting points. Misalignment of the nut to the stud can cause cross-threading, which potentially ruins the hardware. Mounting a tire also demands the accurate torque on a nut: Too much force can damage the hardware, but too little can cause a tire to come off during use, potentially causing serious injury to the passengers. RoboTires’ robot can change car tires, a physically demanding task, precisely, safely, and efficiently. Courtesy of RoboTire. Vision enables a robot to work on two sides of a car to automatically change car tires. Courtesy of RoboTire and Zivid. Passenger car tires weigh between 15 and 25 pounds without the rim. Service technicians must wrestle four of them into place and then perform mounting operations carefully. “People have been changing tires for a long time and it’s a very physical job,” said Benjamin Wilson, senior manager of operations at RoboTire. The company’s flagship product is a vision-guided robot tire-changing system. It is in service at sites in Arizona, Texas, and Pennsylvania, with plans for further locations. The automated tire-changing system uses 3D vision for a variety of tasks, according to Bradley Vargo, lead vision engineer on the project. In the robot tire-changing operation, a technician drives a car into a work area and uses the lift to get the vehicle into the air. After that, the robot takes over, handling the wide variation of tires and cars through proprietary software and vision. “We use a 3D point cloud camera to identify not only where the vehicle is at but also where the wheels and lug nuts are at. In what sort of pattern are the lug nuts? What is the diameter of the tire so we can grip it accurately?” Vargo asked. 3D imaging required If cars and tires were all alike, then a 2D camera might be able to work, because simply getting information on the x and y location could be sufficient, Vargo said. But due to variations in vehicles, wheels, and lug nuts, tire changing requires depth information as well as image data that can achieve under a millimeter accuracy in x, y, and z positions. The only solution is a 3D camera. 3D cameras capture wheel, stud, and lug nut location. Robots use this information for tire changing. Courtesy of RoboTire and Zivid. RoboTire chose the Zivid Two industrial 3D camera for its initial deployments, primarily because wheels and lug nuts are shiny metallic parts and difficult to image. The Zivid cameras have a dynamic range of over 100 dB, meaning that the intensity of image pixels can vary by a ratio of 1-to-100,000. Vargo said the system is designed for standard industrial lighting. Tire-changing bays, though, may open to the outside, leading to lighting differences. The camera hardware handles some of the lighting variation, while the AI software and the remainder of the system in RoboTire handles the rest. Another reason for the Zivid selection is that the cameras produce stable data despite temperature swings. Because tire-changing bays are at least partially open to the elements, they feel the heat of summer and cold of winter. The robot uses other sensors in addition to vision. For instance, it monitors the torque used to put on a nut or take it off. If the torque is too high in either case, it alerts an operator. According to Vargo, RoboTire virtually eliminates cross-threading, which is not a rare occurrence when tires are changed manually. When taking a nut off, the force may be too high due to it getting overly tightened during the previous installation of the tire. The system also has sensors to ensure technician safety. For instance, it detects when someone is in the work cell and the robot stops until the area is cleared. A robot prepares to change a tire, with vision enabling the robot to handle a wide variety of tires, rims, and vehicles. Courtesy of RoboTire. According to Wilson, an initial design goal for the product was to make it safe for operators. It also increases safety for consumers, he added, because it ensures that tire mounting meets specifications and eliminates errors that can arise from manual tire changing. It also saves time, with a throughput of about three cars per hour. That’s about a 50% increase over what a typical shop does. “So, you can increase your production and get it done safer,” Wilson said. RoboTire has future plans for addi-tional features. According to Wilson, it was too early to go into specifics, but one possibility may pertain to using lasers and vision on the system to develop a digital inspection capability.