Most people in the US have experienced the difficulty of driving in snow in one way or another. It has been a constant challenge because the roads become slippery and road visibility is often an obstacle. In line with this, self-driving vehicles need to be able to navigate through the roads during winter better than present models.
Ford, for example, has disclosed six facts about its self-driving technology that will enable vehicles to brave through the snow conveniently. Ford has built high-resolution 3D maps that will scan the area thoroughly before driving through it.
Ford Fusion Hybrid self-driving cars come with LiDAR technology to first create digital maps of the area where the cars will be driven. Using four LiDAR scanners that yield 2.80 million laser points per second, the vehicle will create digital models of the road and the infrastructures that surround it. The map will then serve as a baseline to identify the position of the vehicle when it is in self-driving mode.
This enables the vehicle to locate itself within the mapped coverage later in time when it is already covered in snow. Apparently, the self-driving technology of Ford requires more data in just an hour than the average mobile data used by one person in 10 years. Mapping the environment in high-resolution requires the vehicle to collect heavy and detailed data about the surroundings. This causes the vehicle to collect and process up to 600GB an hour.
Meanwhile, an average person in the US consumes 21.60GB of data a year, amounting to a total of 216GB for 10 years. The LiDAR sensors are extremely powerful. In fact, they can also identify raindrops and snowflakes falling. Because of this, the sensors may return a false impression that there is an obstacle along the way. However, of course, there is no need to steer around precipitation.
To deal with this dilemma, Ford, together with researchers from the University of Michigan, devised an algorithm that can recognize rain and snow. Consequently, they are filtered out of the vehicle's vision, allowing the self-driving car to continue driving. Ford's self-driving cars are built with navigation systems that are far better and more accurate than the average GPS. Present GPS can only be accurate up to slightly more than 10 yards.
Self-driving vehicles need more than that. Ford's self-driving cars can accurately locate themselves within a centimeter; thanks to the powerful technology that scans the environment and compares the acquired data to the stored 3D maps. The data from multiple, high-tech sensors, together with the smart monitoring of sensor health, are capable of keeping the self-driving car out of the blind. Aside from the LiDAR sensors, Ford's self-driving technology also includes cameras and radars that monitor the surroundings using the data from all the sensors fused together.
This process, which yields a robust 360-degree situational awareness, is called sensor fusion. This means that if any elements such as snow, ice, debris, and grime build up on a sensor lens, inactivating a sensor, the self-driving technology remains uncompromised. The camera and radar systems can still identify the deterioration of sensors, keeping them in ideal working order. Moreover, this might even enable the vehicles to perform self-cleaning and/or defogging measures in no time.
Astrophysics major Wayne Williams, the first person to perform a self-driving test in snow, never even thought about being in a self-driving vehicle. Before Williams joined the self-driving team of Ford, he used to work on remote sensing technology for the federal government.
As a self-proclaimed geek, he has always been interested in self-driving technology. However, Williams never dreamt of one day becoming part of a team that will develop the technology and bring it into materialization— let alone be the first to conduct a self-driving vehicle test in snow. According to him, it was all purely business with a coworker monitoring the system from the back seat. Williams said that everyone in the team was confident about the extensive development they had performed and worked on.
However, overwhelmingly, he expressed that it was not until after the test that the true victorious feeling began sinking in. Ford is the first automobile company to conduct a public self-driving vehicle test in the snow. It took place in Michigan, including at Mcity, which is the University of Michigan's 32-acre, real-world driving environment. The testing on the full-scale simulated urban campus is geared toward realizing Ford's mission to further advance the emerging self-driving realm.