UST 031

32 comparable EV, but with the addition of some specific procedures that apply to the autonomous technology. “One of the main advantages we have is that it is connected to our system all the time,” he says. “We have much more information than we would from a standard EV, for which maintainers have to wait to return to the workshop.” The company uses three tiers of maintenance for the EZ10 that are defined by the types and levels of skill required to carry them out. M0 tasks are preventive ones such as cleaning Lidar and camera lenses, for example, that can be carried out routinely by the operator. M1 tasks relate to the platform as an EV, and are centred on replacing mechanical and electrical components that can be carried out by any qualified vehicle mechanic – so long as they are aware of items specific to the EZ10 – as they are more or less standard automotive tasks. M2 tasks are mostly related to programming and configuring equipment such as sensors and calibration of the vehicle to ensure the sensors can see what they need to, all of which takes special skills, Cognet explains. His team is working to streamline M2 tasks at the moment, by creating software tools for maintainers to use. “It’s not programming in the sense of creating code,” he says. “It’s more a case of changing values and updating configurations.” For example, replacement Lidars are not configured when they come from the supplier, so they need to be set up to work properly. That can be done in the workshop or directly on the vehicle. “Configuration could mean changing the field of view or the operating frequency, for example, and setting it up to provide the data that the system needs in the correct format. It could also mean configuring the Lidar’s IP address and its access management,” Cognet says. To avoid the kinds of errors that can creep in when this kind of process is carried out manually, Cognet’s team is developing more automated methods. “The intention is to create a configuration April/May 2020 | Unmanned Systems Technology Stereo cameras replaced monocular versions in the third-generation EZ10’s sensor, supplementing active range measurement from radars and Lidars The current Gen 3’s sensor mix includes Lidars that cover long and short ranges and contribute both to precise positioning and obstacle & collision avoidance For the Lidars, we can check the mechanical positioning, and we have software for fine adjustments to data interpretation