With industry deregulation in the 1980s, the trucking industry has seen drastic changes to the profession. What was once a coveted and well-paid profession, has been plagued by problems that include long hours, low pay, hazardous working conditions, and extended periods away from home. Combined with dispatch problems that take them to wrong destinations and long wait times at loading docks, many truckers are frustrated with a system they feel powerless to change.
Motor carriers that don’t feel they can increase wages and stay competitive, often have a revolving door of new recruits coming and going in constant cycles. As the old school drivers age out of the profession, a new wave of immigrants has largely begun to replace them. The result is a drastic shift in the balance of power that favors large international motor carriers that have transportation contracts with major shippers, while the majority of small carriers depend on overflow loads for which there often is an abundance of lanes with long deadheads or long wait-times.
Stefan Seltz-Axmacher, co-founder and CEO of Starsky Robotics, a San Francisco high-tech startup, thinks he can improve the lives of truckers while taking advantage of driverless truck technology. His approach to HAV trucking integrates robotics and artificial intelligence with old-fashioned trucker experience and know-how. He also thinks that he can scale his take on driverless trucking to commercial success much faster than the bigger companies who are struggling with the complexity and expense of LIDAR-dependent autonomous driving systems.
The Starsky solution? Let the high-tech systems do the easy highway driving, then turn operation of the vehicle over to an actual experienced CDL driver who completes the trip through remote-control operation. For the near term, Starsky will have drivers aboard the trucks, operating at whatever level of autonomous operation conditions and current regulations will allow, with the remote driver essentially operating the truck as a drone for the last stages of the run.
The Evolution of Long-Haul Trucking
Starsky intends to contract with trucking firms for long-haul business with its own drivers. Instead of displacing truckers from the industry, the company will hire them. Rather than sending them out on the road, however, the company will give truck drivers offices, regular hours, and the chance to go home every night. Starsky’s drivers will work at what appear to be simulators, but are actually remote-control cockpits. The remote drivers will earn about the same amount as an experienced long-haul driver.
Because most of the trip will take place on the highway under a high level of autonomous technology, a Starsky remote driver will be able to monitor or “drive” several trucks at once, with hands-on remote control only necessary for the last few miles and in the freight yard.
Starsky Robotics will not depend on LIDAR (light detection and ranging), the technology that many players in the autonomous vehicle industry are using to detect traffic and road conditions. The main problem with LIDAR is its expense. It currently costs about $80,000 to retrofit a vehicle to 360-degree vision capability at high resolution. By using less expensive cameras, radar, and artificial intelligence, Starsky expects to compete effectively on the cost side without any sacrifice to safety. Starsky therefore has the advantage of not needing to wait for LIDAR costs to come down before it can compete in earnest.
Development and testing of Starsky’s autonomous and remote-control driving systems has relied heavily on close cooperation between software engineers and experienced truck drivers. Rather than starting from a technological perspective, Starsky has leaned on its trucker staff to inform and refine the work of the software and robotics teams.
Starsky has successfully tested its technology in both highway and freight-yard operations in Florida, and it has even made some revenue-generating runs while developing the experience necessary to refine its artificial intelligence capacity. Apparently, data inputs from cameras and radar alone can be sufficient for safe autonomous operation, provided that the AI features of the system can “learn” what it needs to know.
For the most complicated parts of a trucker’s work, that is, navigating a city street grid and backing smoothly into a loading bay, Starsky’s remote drivers already know how it’s done. Using the experience of seasoned drivers to fill in technological gaps just might give Starsky Robotics the inside track toward industry leadership in the driverless future of trucking.
