The new frontier of smart cars: remote diagnosis and repair via software

While the concept of a self-repairing smart car seems like science fiction, cars that can be repaired remotely via software are already a reality, as demonstrated by the Xiaomi case. In 2024, the fact that the Chinese tech giant had produced and launched its first vehicle sparked global interest. The model, called the SU7 , is an electric sedan with a sports car design that industry insiders have even compared to a Porsche racing car.

Just recently, the Chinese company announced it was remotely repairing around 116,887 units of its car after a fatal accident that claimed the lives of three college students. Last March, the young women aboard a Xiaomi SU7 were using the active semi-autonomous driving system, Navigate On Autopilot, when they hit a concrete barrier on the highway at approximately 116 kilometers per hour, losing their lives. It's still unclear what exactly happened on board. According to Xiaomi, the car detected an obstacle on a section of highway under construction and issued a warning before handing control over to the driver , who may have been unable to avoid the impact.

The social media controversy, but above all the intervention of Beijing's state authorities, have forced the Chinese giant to intervene. Xiaomi has thus announced that it will remotely fix a defect found in the assisted driving system of over 110,000 SU7s , a third of those on the domestic market. According to the state market regulator, the assisted driving system on highways has demonstrated insufficient recognition, warning, and management capabilities in complex driving conditions. "Under certain circumstances, when the driver assistance system is active, the system may not be able to adequately identify, warn, or respond to unusual road conditions." This is the official statement regarding a defect that could increase the risk of collisions if drivers fail to intervene promptly. And that's exactly what happened.

But how is it possible to repair a car remotely, without even physically touching it or performing a visual inspection, like when we take our car to the mechanic? Does this mean that in the near future, roles like mechanics and mechatronics technicians will no longer be needed? The most plausible answer is that they will continue to be indispensable for all those tangible breakdowns: worn brakes, tires in need of replacement, batteries in need of a change —in short, the classic problems. Except that the automotive world is no longer classic; it's changing rapidly with the transition to electric vehicles, and software is increasingly becoming the central focus of vehicles.

Today, remote diagnostics and intervention are possible when the problem isn't mechanical, but rather software-related. Many on-board systems, such as those that manage the engine, brakes, or infotainment system, are controlled by software that, like any other program, can suffer from bugs or malfunctions. In these cases, car manufacturers can send updates called OTA, short for "over-the-air," as is the case with any other electronic device, such as a computer. Smartphones. Tesla has also already implemented advanced remote monitoring systems that allow technicians to analyze vehicle data and, in many cases, resolve the issue with a simple software update.

These are data packets transmitted directly to the vehicle's computer connected to the network, which can fix programming defects, improve performance , or even add new features. All this happens without the driver having to take a single step to the workshop. This is the case with the repairs on the SU7. And that's not all. Thanks to the implementation of artificial intelligence in vehicles, predictive maintenance is developing ever more rapidly, and this is the next step in automotive diagnostics, allowing for earlier detection of anomalies and timely intervention. Moreover, with the amount of data coming from the vehicle and analyzed by automakers' computers, the time cannot be far off when we will know in advance if a sensor is about to fail, an ADAS safety function is compromised, or perhaps even an electronic or even mechanical part . Moreover, if AI is able to calculate the probability of a given factual event with increasing accuracy, it could be possible to predict car problems in the future. For example, software can detect an anomaly in a battery cell before it causes a serious failure and the car stops working.

While Xiaomi was forced to intervene remotely after a tragic event, other companies are developing remote repair and diagnostic projects. For example, Hella Gutmann Solutions offers workshops access to a remote diagnostic service for vehicles from around 70 different brands. A technician in the workshop connects a VCI (Vehicle Communication Interface) and a dedicated device to the vehicle and the internet can perform various interventions via software. And the path being paved for electric cars is increasingly " software-centric," meaning that much of their performance, functionality, and even battery life depends heavily on the software and electronic control units.

This means that many problems that in a traditional car would require mechanical intervention in an electric car, remote diagnosis and intervention are the ideal tool for reading error codes, analyzing sensor data and identifying the root of the problem.