I need to repurpose Stanley Kubrick’s classic Dr. Strangelove subtitle to “How I learned to stop worrying and love the sensor.”
A 1916 Fageol farm tractor engine I viewed recently might have had one sensor: the farmer. Faegol Motors Co., based in Oakland California, dabbled in making cars, buses, and trucks in addition to its primary tractor products. Fageol later became Peterbilt. You get points if you already knew that.
Today’s semi-trucks seem to be growing sensors like mushrooms. The implementation of engine controller ECUs and ABS braking systems in the 1990s really began the waterfall growth in software and sensors in the trucking industry.
Software codes that control nearly everything on a modern truck require a lot of data. That data comes from sensors. Major enablers of this communication are common databuses like the venerable SAE J1939 standard titled Recommended Practice for a Serial Control and Communications Vehicle Network.
Introduction of Advanced Driver Assistance Systems (ADAS) has brought in radar, lidar, camera, GPS, and other sensors supplementing basic operational sensors like those buried in engines, transmissions, and braking, and heating, ventilation and air conditioning (HVAC) systems. We have low air pressure systems now monitoring tire pressures.
Every new model year seems to include more sensors. Autonomous vehicle (AV) systems likely will add even more sensors in the endless effort to supplement or replace human senses.
The challenge facing the trucking industry is that each of these additions is adding lines of software code to an increasingly complex system of controllers. Fully testing all the permutations of potential failures and symptoms is a Herculean task. Maintenance people need to be familiar with this ever-increasing complexity, continuously training and updating test equipment and procedures.
An automotive example from my own experience
I had a vehicle where the tire pressure monitoring system (TPMS) was spuriously signaling a loss of pressure in my tires when I was on the highway. I would find a safe spot to park off the road and manually check the tires, only to find they were all properly inflated. I took this car to a certified factory dealership to diagnose and repair. It turns out that the TPMS sensors have an internal battery, and I had owned the vehicle long enough that those batteries were at the end of their life — the batteries were not dead, but erratic. That analysis had my car tied up for two days. It did not stop there. The shop went to replace the sensors in the tires and found the car would not talk to the new sensors. The car stayed in the shop for several days as they struggled to figure out the problem. Eventually the shop called the OEM and determined the shop equipment was not on the most current version of software needed to update the car to the new sensors. All in all, I think my car was tied up for a couple weeks over what seems now like a pretty trivial problem.
I expect there are many examples of truckers pulling into service bays where the problem takes more time to troubleshoot than they have patience for. Downtime is one of those interesting metrics as drivers are not typically paid when the wheels aren’t turning, companies are not making money when the freight is not moving, and delays in delivery are likely also impacting the shipper. The service personnel and the shop, however, likely are being paid for every minute the truck is in their hands.
One industry source estimates standard semi-truck shop labor rates are $195 to $300 an hour for initial diagnosis. Rates on the actual repair time vary.
The Bureau of Labor Statistics estimates that the median pay for driving a semi-truck works out to about $23 an hour. The fleet might be penalized for late delivery, it might lose future loads, and it may have to dispatch another truck and driver to handle a stalled load. A late delivery might slow down a shipper’s production line. The ramifications of downtime can be significantly costly to the driver, the fleet and the customer.
To the best of my knowledge, those downtime costs are not factored into tracking systems. What is tracked are service labor, parts, and warranty. OEMs and dealerships do track vehicle downtime, but I do not believe they apply any financial ramifications from the driver, fleet or customer perspective.
The increasingly complex vehicle world is starting to exploit sensors to help reduce diagnostic time. Automated diagnosis might even be occurring while a truck is on the road, predicting future failures and diagnosing root causes so that the vehicle can be proactively repaired quickly at the next service opportunity.
There is huge potential for real-time smart analysis of the massive amount of data being produced on trucks. That analysis will be done by evolving Artificial Intelligence (AI) systems. Some of the analysis may be accomplished by the vehicles itself, or by cloud-based systems monitoring the vehicle.
AI systems will get better in time, just like humans gain experience. They likely will discover that they don’t have the right data or sensors to improve, which will lead to adding even more sensors to the trucks. More complexity. More failure modes to deal with since sensors themselves also can fail, like in my TPMS tire sensor.
Who benefits most?
I’m convinced that two careers are going to be very profitable in the future. The first is skilled service technicians. They will have to constantly be learning, but they will be indispensable as the tip of the sword for AI service information validation. The second future career will be in sensor development. Companies that develop sensors will be under relentless pressure to make sensors more reliable, more accurate and less expensive.
A parallel industry to look at is aircraft development over the last century. The evolution of aircraft has seen huge growth in sensors, and the service knowledge needed to keep them functioning.
I expect the complexity of trucks will continue to grow, but I have faith that the sensor developers and the service community will keep up. They’ll have to. We shouldn’t worry.
