The next time a street sweeper finishes its run, take a look at the litter it left behind for another pass. The problem is that drivers must not only navigate the truck, but also simultaneously control the brushes to apply the optimal force for the type of trash on the road. A team at Surrey University suggests that drivers should put some of the responsibility in the hands of an automated system.Researchers Graham Parker, Maarten Michielen and Gareth Peel of the university's Mechatronic Systems and Robotics Research Group believe that semiautonomous street sweepers could offer faster cleaning speeds, lower fuel consumption and lower emissions. The sweepers would feature a system that adjusts the brush speed and pressure to the road conditions and litter ahead, ensuring that only one pass is needed to clear the surface, and that the brush is kept close to the curb, where 80 percent of trash is found.As conceived, the system will project a laser line ahead of the vehicle. A digital camera will monitor the shape of the line to locate and classify types of litter by their outline. The brightness and color of the litter also offer clues that the system could use to distinguish, for example, wet and dry sand, which require different brush settings.The concept system currently relies on image-processing software to discriminate the type of trash, and a decision-making controller selects the appropriate axial force for the brush from a look-up table, Parker explained. He said that future iterations could be equipped with neural networks so that users could train the system to handle different conditions, rather than relying on predefined variables.Parker said that although the researchers have no plans to commercialize the technology, they have attracted the attention of the industry. Nevertheless, he feels that smart sweepers will not appear on the road anytime soon. "One of the problems," he said, "is that the UK does not have enough manufacturers of this type of equipment."