GM has been working on its Cruise driverless cars for a long time and back in March, the first grumblings began to surface that all wasn’t well with the Cruise project. In March, reports indicated that the Cruise cars frequently swerved and hesitated due to difficulty identifying objects on and near the roadway. A new report has surfaced that cites one current and three former Cruise employees who have witnessed issues with the project.
According to those sources, the Cruise cars still struggle identifying if objects on the road are moving or stationary. This inability to identify if something is moving or not leads to the vehicles hesitating or stopping when the Cruise cars attempt to pass a row of parked motorcycles or bicycles. These sources also claim that the software that runs the autonomous cars is failing to recognize pedestrians and has braked erratically at times because the car sees a bicycle that isn’t there.
Erratic performance and the inability to identify objects correctly aren’t the only issues GM faces. Cruise has yet to enact the data-sharing collaboration with the San Francisco Fire department to prove the cars can respond to fire truck sirens; this is a necessary step to train the cars to react appropriately when on the roads.
Another issue has to do with the open-source software robotics tools that Cruise is using to develop its tech, that software reportedly has delays that slow messages from the sensors of the car from reaching the car’s brain. Cruise CEO Kyle Vogt has claimed that the company’s next generation hardware and software will address these issues.
Vogt maintains that he is “sure” Cruise was building systems that didn’t meet requirements early in the testing and development process. Another potential roadblock for Cruise is that it must meet certain performance targets to get the money Softbank and Honda have promised to continue development and operations. Softbank has promised $2.25 billion in funding for Cruise.