5-19-17 Dragline

In Dragline’s Bucket

Glowing predictions about the entry of autonomous vehicles on our country’s highways and streets strike us as too optimistic. Especially given the murderous mental state of too many human drivers. We also feel the same about autonomous construction equipment which, superficially, seemly harbor a better potential. One would think a construction site could be controlled far better than the open road, with all its tailgaters. Yet, as we observe, when the equipment shows up the problem won’t be found in its circuit boards and hydraulics. We should seek it in its programming. Our concern is the people writing the code will think they know all about the functions of construction equipment when, in fact, they’re relying on too many stereotypes. How much unnecessary financial weight does duplicate regulation of construction cost our nation? It appears to be quite a huge figure, according to testimony in the U.S. Senate earlier this month.

Autonomy Isn’t King

Over the past couple of years we’ve read rather shallow accounts about the development of autonomous construction equipment. For want of a better term we call such stories “glossies.” They’ve appeared on web sites, and the few “equipment books” that are still printed. Typically they’re pitched to appeal to their publication’s advertisers.

A primary reason we call these stories shallow is, other than discussing available equipment and hardware, they slide over the human factor. They may mention, almost in passing, the unique difficulties that can be encountered on construction sites, but their attention is drawn to product successes where site environments are easiest to control.

One of them has been surface mining. Komatsu has been a big player in this field, using autonomous models such as its 930-E electric drive hauler to move mined material on defined routes, with sensors that keep track of payloads, trips, the hauler’s maintenance records, and need for repair. Yet keep in mind a strip mine is a controlled, regulated site that’s supposed to be off limits to the unauthorized. Also, an autonomous hauler is intended for one function. And, perhaps as a bonus, on a hot day a robot won’t sneak off for a cold beer.

As a published science fiction writer (albeit that happened a long time ago,) we’re not against intelligent technology, artificial intelligence, and autonominity. In our lifetime we expect to see autonomous construction equipment, including its programing and monitoring, as a component of building information modeling. The two will interconnect for efficient construction. But, for safety’s sake, we think it will have to include input from the worker we mentioned in the above paragraph, drinking a Schlitz.

The inspiration for what some may think is a strange line of thought was sparked by a “radio” glossy about research underway at Volvo for the development of the autonomous garbage truck.

We have expertise in the collection of garbage having spent about half a year doing so for the city of Detroit while earning our way through Wayne State University, in the spring and summer of 1972. It was great therapy for a young man with a broken heart.

Some foolish people think that, because a garbage truck has (but rarely travels) a predetermined route, with predetermined stops, (that may not always be possible to make,) it can be easily developed into an ideal autonomous vehicle.

Well yes. And no.

The radio commentary cast this ideal vision of an autonomous garbage truck moving down a residential street with a sole operator walking by its side. It would stop, the operator would load its hopper, and it would continue on. The sky would be blue, the birds would be singing, and a friendly squirrel would offer up his nut.

Simple? Yes. Realistic? No.

Let’s ignore safety concerns and assume autonomous technology will be sophisticated enough to always detect the child chasing a errant ball into the street, or a backing up SUV that somehow misses detection, or the tiny dog that got away. Not to mention how its systems would react and slam on the brakes in time to stop a garbage truck that weighs 33,000 pounds empty and approaches 51,000 pounds when full.

When we were working for the city of Detroit we had a truck driver who stayed in the cab for the most part and a two man crew filling the hopper. Now, somedays we’d be assigned a residential route and somedays a commercial one. Two people working the hopper were necessary because one could easily become injured by the heavy weight of some of the objects we were required to lift, especially if they were saturated by rain. Two were also necessary for meeting our production quota — you weren’t given credit if a complicated pickup slowed you down — there were times we raced down streets and alleys to get back on schedule.

Today many cities contract with commercial firms and typically they use a truck with one driver and one person to load. (We’d rather not look at their worker injury rates.) With an autonomous truck you’d only have a loader, and he or she will quickly tire out from the physical exertion of loading, if not being soaked in rain or freezing in a blizzard.. Yet, as they drag along, trying to keep up with their autonomous truck. do you think they’d do a good job making sure it was functioning properly? Enough to manually stop it if an unexpected danger happened? Something the truck’s programing missed?

We’re not saying this problem won’t be solved. Only that we anticipate a long learning curve. It’s a curve that may be scoffed at and neglected by technicians and senior management who’ve never manually performed the work. Without giving much thought to it, they will be programming using stereotypes of worker behavior and the situations they will face.

Often we’re forced to use stereotypes. Our understanding of what happens in an operating room essentially is based on television dramas because, when we’ve been in one, we’ve been heavily sedated. The same is true for most professions. Until you gain “hands on” experience in an occupation your expectations of it have a high probability of being unrealistic. Thus, if you’re given an assignment of doing something that impacts something you know little about, you can fail because you’re trying to impose a stereotype that’s unrealistic.

Stepping beyond technology snafus, this type of mistaken thinking has continually plagued our industry. It crops up in hiring, where job qualifications are defined inaccurately, leading to the hiring of individuals whose performance doesn’t match company needs. Occasionally it causes a mismatch in equipment selections. Or it masks a glaring error by using an extravagant and unnecessary investment in equipment to cover incompetence.

During the last mid-century and through the 1970s the use of helicopters in construction was both eye-catching and trendy. We prepared feature stories about one contractor in particular who’d take us from construction site to construction site in his Model 206 Bell JetRanger II, which today can be purchased for about $900,000 to $1.2 million.

Helicopter fueling and maintenance is expensive but that didn’t deter the contractor. It gradually became clear to us, however, his principal reason for using it was to provide face-to-face management of all of his projects within a single day, making decisions and performing functions typically delegated to project superintendents. He may not have realized it but he was communicating mistrust. Either his middle managers were incompetent, which meant he should have replaced them, but more likely his ego was getting in the way of his leadership.

His company breathed its last in the mid-1980s, stuck in a “problem project” in Florida.

Back to today, over the next five years we expect autonomous construction equipment to be introduced. Probably it will first show up in the earth excavating market but its reach will gradually broaden. It will compete with semi-autonomous models, where an operator standing by can get it to autonomously perform tasks by entering commands into a smart phone.

Anticipate a slow learning curve. There will probably be a period of time where manual operation will be better and faster than autonomous. We can’t predict how long that period will be.

But remember, in the coming autonomous age, how the equipment is programmed will be far more important than its mechanical capabilities. Hopefully those who develop this technology won’t make the mistake of ignoring the advice of expert construction equipment operators.

The Big Bog Down

Because the federal permitting process doesn’t draw much attention from the general news media, the public doesn’t understand how unnecessarily time consuming and expensive it is. Of course it’s irresponsible to build anything without appropriate reviews, especially of environmental impact, but the process cries out for speeding up, primarily through the elimination of redundancies.

This point was hammered home last week by testimony delivered before the U.S. Senate Committee on Environment & Public Works by Leah F. Pilconis, senior environmental law and policy advisor for the national office of the Associated General Contractors of America.

It came out early in her 36 page presentation. She’d found a source that estimated the financial impact of the problem.

“[A] National Association of Environmental Professionals (NAEP) review of the 194 Environmental Impact Statements (EIS) published in 2015,” she reported, “found that the average time to complete an EIS was five years and only 16 percent were prepared in two years or less. Meanwhile, a 2015 report by Common Good, a non-profit government watchdog, finds that a six-year delay in starting construction on public projects costs the nation more than $3.7 trillion in lost employment and economic gain, inefficiency, and unnecessary pollution. That is a staggering amount of statutory and regulatory inefficiency that needs to be addressed.”

She called the time “ripe” for Congress to do something about it. There seems to be way too many layers of bureaucracy bogging down what should be straightforward.

“The current process of performing sequential and often duplicative environmental reviews and permits on the same project – performed by all levels of government following the National Environmental Policy Act (NEPA) approval process – is presenting massive legal hurdles to infrastructure approvals,” she said. “A builder of infrastructure—whether a contractor or government agency—must seek approval not from ‘the government,’ but from a dozen or more different arms of the government. According to bonding companies that finance large public works projects, two environmental approvals are critical in rating a project’s risk for bond financing. Those are the NEPA review (1,679 days, on average, to complete an EIS) and Clean Water Act (CWA) Section 404 permit authorization (788 days, on average, to obtain an individual permit.) Obtaining these approvals prior to bonding greatly reduces risk and achieves a higher bond rating to the benefit of the project sponsor.”

In an earlier Dragline we recorded our doubt President Trump’s proposed $1 trillion infrastructure investment program can get off to a quick start, using “shovel ready” projects. Ms. Pilconis has documented another major reason why.

May 19, 2017