How to fix a car � without a mechanic Think of a mechanic working on your car, and youll probably imagine oil, grease, and an array of complicated tools. But the workshop of the future, for either the professional or home mechanic, could look much more hi-tech. That because modern vehicles are becoming more and more complex, with a vast array of components. Whereas in the past a mechanic, or even car enthusiast, could reasonably recognise and understand everything that made an engine run, now the proliferation of sensors, computers, and safety devices makes this near impossible. BMW vision for the way its engineers should be working on its cars is different. Instead of reaching for any of the shiny silver tools on his cart, the mechanic picks up a pair of what looks like sunglasses with connected buds for his ears. He glances back over to the engine, and this time he sees each component highlighted in bright colours, and is given computer-generated instructions on what to disassemble, in what order. The glasses the mechanic is wearing contain small screens to provide an overlay of computer-generated images on the real world augmented reality (AR). From Google Glass competitors, to contact lenses with built-in displays, this tech is coming in part enabled by the wide adoption of smartphones. It seems to be almost waiting for people to come up with novel applications for it. BBC Future has reported before on the future vision of cars with augmented windscreens, or even see through backseats, effectively making the car seem transparent to the driver and helping visibility. Most of those concepts wont be reality until far off in the future, but car companies are jumping into the augmented reality world. Volkswagen is using AR, and promoting it as the futuristic way to provide servicing for its new concept, the XL1. The Augmented Reality maintenance and service support system is named Marta (Mobile Augmented Reality Technical Assistance) and will provide instructions to help identify and label vehicle components. We see a lot of use in the automotive industry. For us it always been a sector which sees the value in the technology, and it always been easy to drive results within that area, says Trak Lord from Metaio, which claims to be the world leader in augmented reality. Smartphone repair kit? As well as car companies using the technology internally to help engineers and mechanics, it is increasingly being made available to the rest of us. Audi has recently released a mobile phone app in Europe that can identify over 200 elements on the inside of a car. It is a user manual, and then some, with the ability to pick out what the camera is being aimed at, and deliver more information on it. Augmented reality apps can also be used for maintenance at home, perhaps making simple fixes on modern cars more accessible to more of us. How often have you heard people remark that older cars are easier to work on, and that the engine of most modern cars mostly just looks like a slab of plastic? AR could help see through that plastic. Say youre driving down the road, and a light comes on you freak out and pull over. You can then load up that scenario, based on the symbols, in an app, says Lord. It could then tell the driver that, for example, it the low coolant light, and to open the engine compartment to check the level while giving instructions at every step. The driver would simply point their smartphone camera in the general direction of the problem. It scans the engine, recognises the engine, and you just see an arrow directly on the area where you are supposed to refill the coolant, says Lord. But the wide adoption of these applications, which overlay information on reality, creates some unique challenges. One of the most important aspects of any AR system is the ability to track image, space, environments, objects, etc, says Lord. That requires several discrete steps, each of which has its own complications. First, the powering up of the camera, where the device has to recognise its orientation and position. Then the camera needs to lock on to, or track, some sort of distinguishing markers, in order to figure out what it is looking at. Then it needs to follow those markers even when the camera is moved around. The system also has to work in multiple lighting conditions. If a customer breaks down at the side of the road at night, in the rain, a reliable AR app on their phone still needs to be able to identify that it is looking at an engine. It involves skills that the human brain is very good at, and can complete almost instantly. We recognise edges and shapes and can instinctively identify them. Computers have a much harder time. Is a straight edge a part of the car, or part of the building it is parked near? The hope is that AR systems will become robust enough that explicit instructions are no longer needed. If a mechanic or a driver has a problem, they will simply turn to their phone, or tablet, for help, which will replace complicated paper-based instructions. To couple [AR] with the user manual is the next logical step. It kind of surprising we dont see this more than we do already, says Dr Jeffrey Miller, at the University of Southern California Computer Science Dept, and member of the Institute of Electrical and Electronics Engineers (IEEE). So the car of the future may come with one simple instruction DONT read the manual!