Autonomous Vehicles – The Big Picture

The Need for Common Regulatory Environment

One major thing to keep in mind with autonomous vehicles is that the goal is (or should be) to bring a greater degree of safety to our roadways, reductions in fuel costs, lowering traffic congestion and improving the environment. One large part of the puzzle in keeping the technology moving forward is that all levels of government need to come together and bring the regulations up to speed with the technology. These systems will not have their fullest ROI if some jurisdictions allow them and others do not – especially if you get jurisdictions that chop up any continuity between regions. An example would be if West Virginia allowed a technology on their roads while it was barred in Pennsylvania, Ohio, Kentucky, Virginia and Maryland. Such a scenario will provide a benefit to the state of West Virginia but it is lower than it would be if the other states allowed the same technology. With the above being stated, it appears there is some progress being made on this front to consolidate regulations at the Federal level: The US is speeding toward its first national law for self-driving cars.

Further, there should be a consistent set of regulations not only between states but also between the United States, Canada and Mexico to ensure the safe and efficient flow of trade between the two countries.

Drivers – A New Opportunity

The reality is that a human driver is not going away any time soon. Even if it is only for liability reasons, the need for a “bum in the seat” will remain so that there is a human that can take control in certain environmental situations where the technology still has problems “visualizing”. For example modern lidar and radar systems will allow for the potential of allowing commercial vehicles to have a greatly reduced gap between vehicles. This can provide significant fuel savings, similar to what race car drivers do as they wait for an opportunity to make their move. However, under very sunny conditions some of these systems have difficulty in distinguishing large white vehicles. So that trailer in that it is trying to follow closely sometimes can be difficult for it to “see”! The radar system can help reduce that “blind spot”, but what about a snow storm or a heavy rainfall, two conditions that can also be difficult for either system to work efficiently? These are cases where the vehicle will need to indicate to the driver that they need to take back control of the vehicle to maintain an acceptable level of safety.

Another situation where a human needs to take control is in stop and start traffic. In these cases, it is the sudden decelerations that are difficult for the automation to handle. So one can think of it this way, on a Detroit to Miami run, the truck could control itself for much of the trip down I-75, but the driver would need to take control through the construction zone between Dayton and Cincinnati as well as through the traffic congestion around Atlanta. Finally the driver will still have to be in control in the final delivery section.

Where there is an opportunity for the drivers is to become more like the captain of a ship, ready to take control when needed but able to work on other tasks when the vehicle is managing itself. Depending on conditions the driver could be scanning and sending their paperwork, sending documents to a customs broker or looking for their next load if they are an owner-operator. He could be able to have the time to prepare his own meals with fresh ingredients instead of eating at a truck stop. She could be using in cab gym equipment to do a workout to keep in shape, reducing fatigue and maintaining better health. Finally the driver could be taking training sessions while on the road making it more efficient and timely to keep your driving force up to date with the ever changing regulations or taking courses towards college credits. By giving the profession a more varied work experience we may be able to attract from a wider pool of new recruits.

At the same time we may be able to get modifications to things such as lane departure systems (that take over control of the acceleration, steering and braking systems) to allow for something like back-up assistance. This would take some of the risk out of introductory drivers who have a difficult time with backing up trailers. It could even help out experienced drivers when they have to blindside into a tight dock. The system could be either active (with the driver allowing the vehicle complete control to back up the trailer) or passive with the system watching what the driver is doing and ready to intervene if they start veering into trouble.

Actively Changing the Public’s Perceptions of Trucking

The last area that is a very real threat is the public perception of autonomous vehicles. At the present time there probably are not a lot of drivers who would be comfortable sharing the road with a self-driving vehicle. There are just too many passenger vehicle drivers who really aren’t that comfortable with commercial vehicles being on the road with them at all. So yes, there is a threat that some lawmakers could try to use these technologies as a wedge issue in an election. However there is an even greater opportunity for the industry to use these new technologies as tangible proof of our commitment to the safety of all drivers on the road. For example, there are new collision avoidance systems that can assist a driver in avoiding a rear end collision. The driver might initiate the turn to avoid the hard braking or stationary vehicle in front of them but then the system takes over, looking at what is to both sides of the truck, the speeds that the other vehicles are traveling and what the distance is to the vehicle in front of them. There really is no reason that the truck itself could not see the hazard and start determining the evasive action before the human driver can even process what they are seeing. By playing up the benefits of these systems to the general public we have a huge opportunity to raise the popular opinion of the industry in general with the add-on benefit of then increasing the attractiveness of trucking as a career. Increased safety, the potential of attracting people who would not otherwise look at driving as a career, and lowering the risk profile of our fleets, and not forgetting the opportunities for increasing profitability.

Blockchain Part II – Trusting the Process

Two weeks ago, we took a look at what a blockchain is.  This week we will build on that base and will start to take a look at potential uses for the transportation industry.

One area that we did not go in depth on was how does a blockchain encrypt data.  Let’s look at a simple example based on one that recently appeared in the Economist.  The data from each transaction in the blockchain gets put through a cryptographic “hash” function that will distill it down into a string of numbers with a fixed length.  This will create the “header” to the data that will be stored.  This “hash” is a one-way street.  While it is relatively easy to go from the data to its hash, it is next to impossible to go from the hash to the data.  What this means to us is that if you make one, seemingly insignificant change to the data, this hash function will return a different value.  This is how any unauthorized changes get noticed.

Let’s back up a bit.  A cryptographic hash function is basically just an equation that takes the data and uses a formula to turn a specific transaction set into a number.  This could be as simple as A=1, B=2, C=3, etc.  Alternatively it could be much more complicated, rearranging the alphabet and numbers, giving values for special characters, etc.  Once the equation has been done once, the program will look at the length of the result and determine if it is the right length.  If it is too short, then it will get padded with some additional zeros.  If it is too long then the result will go back through the equation to get reduced (similar to when we all did reducing fractions back in grade school).

Take a set of 4 transactions – for our purposes, consider them to be 4 different pro bills.  Each pro bill has a shipper, consignee, piece count, weight, and rate.  The data on each pro bill are different.  A blockchain will run that data through a mathematical function to create the “hash” value of a length specified by the application – let’s say 16 characters for this example.  Pro bill A will be put through the function and get a unique hash.  The same will happen with the other 3 pro bills.  Now the program will take the hash values for pro bill 1 and pro bill 2, run them through the function again to get another unique hash value.  The same will happen with pro bills 3 and 4.  This system of combining is called a Merkle Tree.  Now we take these two hashes through the function again to get a final hash that will get put into the header along with a date stamp and a special value called a nonce.   For now consider the nonce to be a random value that helps to make cracking the security difficult and time consuming.

For a very simple example, let’s assume that the data set only includes numbers.  Additionally the equation only adds the numbers together with the hash value of the block before it and then ends with a count of the numbers involved (example – 983 would give a count of 3). The hash needs to have a length of 8 characters.  A further simplification is that we will not use a date stamp or a nonce value – they are not necessary to see what happens.  The first transaction is for 234 pieces, 2 skids, 4213 pounds and a declared value of $8000.  The equation will calculate this as 0+234+2+4213+8000 = 12449.  The count is 12.  So far we have 1244912 which is only 7 characters long.  To get to a length of 8 we will pad this with a single zero at the start to get a value of 01244912.  Now let’s say we have a revision and this now weighs only 4210 pounds.  This goes through the addition to get a value of 12446.  The count is still 12.  The result is still only 7 characters so the padded result is 01244612.  As you can see this does not match the original hash, indicating that something has been changed.  The header does not tell you what has changed only that something does not match.

Now that this first block has been entered into our chain, the hash value from this first block will serve as a value that contributes to the hash of the second block.  In our example we then would have a first value of 01244912 and then add the other pieces of information to it, do the count and then see if the length is ok.  What this means is that any changes to previous data blocks will also be exposed.  The big “so what” is that this creates a secure and transparent ledger that will allow a member to follow a chain backwards and be able to easily note where changes have occurred.   You will still have to examine the records to see what has changed.  However if the values match then you can trust that things have not changed and that you can trust that block of information.  In effect other parties have “audited” the transaction and also agreed that it is the same, something that you are likely (currently) paying someone to do.

Next week we will look at more practical examples of what a blockchain can (will) do for your trucking business.