ChainLink and MIT Enterprise Forum co-organized and hosted a panel session at MIT recently: IoT Takes to the Road: Getting Smarter in Transportation. The outstanding panel included CEOs and CTOs from some of the most innovative firms providing IoT in transportation solutions (see sidebar ‘The Panel’).
Marrying IoT Data with Broader Data Sources
In their introductory comments, panelists talked about the IoT data used in transportation and how modern connected trucks generate hundreds of types of data, such as engine RPM, vehicle speed, acceleration, location (via GPS), fuel used, hard braking, coolant temperature, oil pressure, and much more (see IoT in Trucks). Airplanes, ships, trains, and other vehicles similarly generate a tremendous amount of data. All of the panelists pointed out that this vehicle-generated data is combined with various external data for different applications. This could include static data like maps and geo-fences to alert when vehicles have strayed from the planned route; enterprise data such as order and shipment information to understand what is on the vehicle; and near real-time data about events and conditions in the environment such as weather, traffic, social media, news feeds, and so forth. (For more on the types of data and how they are used in logistics, see The Journey to Visibility).
So what are companies doing with all this diverse data?
Improving Fleet Performance
Both Fleet Advantage and Descartes use that data to track driver safety, fuel consumption, fleet KPIs, and vehicle/driver performance for the fleet owner or lessee. This can help them continually improve fleet performance, spotting and resolving equipment or driver issues. In addition, Fleet Advantage does an economic analysis of the potential performance and benefits vs. cost of moving to newer vehicles. They use that to advise fleet lessees about when they’ve reached the tipping point, the optimal time to switch to a newer truck with better fuel economy and safety features. By combining this with financing innovations of shorter lease periods (3-4 years instead of the typical 5-7 years), they allow tier 1 lessees to swap out vehicles sooner, bringing those vehicles into the tier 2 market several years sooner than they would have been available traditionally, giving all the tiers greener, safer vehicles.
Providing Visibility and Exception Monitoring
Descartes, Savi, Weft, and Transvoyant all use the data to provide their customers — shippers, consignees, carriers, 3PLs, freight forwarders, or other authorized interested parties — with visibility into location, condition, and status of in transit shipment of goods. Some of them also provide analytics with rules and machine learning that combines the IoT data with other data, such as weather, traffic, and events, to more accurately predict the precise arrival time of goods.
What Happens When Our Washing Machines Start Ordering Their Own Detergent?
The first question posed by moderator Ann Grackin was about the challenges that will arise when devices in the home or workplace start ordering their own supplies — using the example of the washing machine that orders its own detergent. This is not something speculative about the distant future — Amazon’s Dash Replenishment Service is already doing exactly that with connected devices (shipping or soon to ship) such as a Brita Water Pitcher that orders more filters, Whirlpool washing machine that orders its own detergent and other supplies, printers from Brother and Samsung that order new print cartridges when they are getting low, as well as self-re-ordering coffee makers, soap dispensers, pet food dispensers, and much more. This trend could dramatically increase the number of individual orders and small shipments to homes and businesses. Ann asked what will be the impact on the transportation industry and how will it respond.
Delivery by the ‘Kid in his ’93 Mustang Making Money on the Side’
Jim Griffin said, it doesn’t have much impact on the long haul or regional carriers who “deliver products to a certain point in the chain. Rather, it is the last mile challenge that has to be solved.” Ken Wood pointed out that a lot of similar changes are already happening because of the new omni-channel retail models which, as Ken said “aggressively challenge the ‘go to the store to buy your stuff’ model of shopping.” He pointed out that they are already working with a lot of fleets and retailers that are doing home delivery today, including white glove service.
Part of this question involves whether these self-ordering machines will need same day delivery. In theory, these machines could give a longer lead time notification of need than today’s consumer who often wants same day or next day delivery. The longer lead time enables suppliers and retailers to more economically combine and schedule the deliveries. But when you want lots of same day or ‘within hours’ delivery it changes the delivery model and what makes economic sense a lot. One panelist pointed out that people are looking at services like UberRUSH, Amazon Flex, and Deliv which change the equation from an expensive truck and driver to “some kid in their ’93 Mustang with their iPhone that just wants to make some money on the side.” What are the business models and intelligence that make that approach economical? There needs to be intelligence in the dispatching to meet service obligations while controlling costs, rather than simply assigning each new order to the first available driver.
Dennis Groseclose said you could “put sensors on everything in my fridge, sense when replenishment was needed, and then gang it up with everything else needed in my neighborhood to optimize deliveries.” He brought up the privacy question about whether or not you want to share that information (more on that below), but added that Amazon already knows what we buy and has invested in lots of intellectual property to figure out what else we will buy.
An audience member asked about the ‘last foot problem’ — what to do when no one is home after several delivery attempts; the issues of leaving items unattended on someone’s doorstep. In response, Marc Held mentioned the use of lockers such as Amazon Locker which lets customers pick up at various convenience stores or UPS Access Point lockers. An audience member pointed out that Volvo has experimented with their Roam Delivery Service, which enables groceries and other packages to be delivered to your car. The car’s GPS tells the logistics provider the location for the drop off and a one-time ‘digital key’ gives the delivery truck driver access to the car; as soon as the goods have been put into the car, the digital key disappears.
Ken Wood said that there will always be people who still want delivery at home, so one solution for them is to offer much tighter delivery windows. That approach has been shown to drastically reduce missed deliveries. He said retailers have been able to up-charge for narrow time-window delivery service. He added that so far this is not available from UPS or FedEx, but that is starting to change. And pointed out that in markets that are dense enough, like say downtown London, they are toying with deliveries to wherever you are at the moment, like at the Starbucks you hang out at — though it is still to be seen whether that kind of service takes off.
Data Ownership and Privacy
There was a discussion of ownership and privacy regarding all the data generated and used in logistics and delivery. It was pointed out that in the EU, they believe personal data belongs to the person, whereas in the US it is the ‘wild west,’ though that battle may not be over yet. A younger panel member said, “The value is there in sharing your information. If Amazon offered a significant discount in exchange for viewing what is in my fridge, I’d do it.” Another said that in the EU you can make that choice, but in the US you give up that choice by agreeing to a lengthy privacy statement with a clause buried in fine print somewhere. The jury is still out.
Someone asked about ownership of the data and sharing with insurance companies. Ken said it varies depending on the provider of the service. In Descartes’ case, the data belongs to the fleet owner and Descartes doesn’t take any ownership of the data coming off the truck. However, there are cases in the consumer world where the consumer can voluntarily share their machine-collected driving data with insurance companies in exchange for the possibility of a discount on premiums.
Liability — Do You Really Want All That Visibility?
An audience member asked about the liability associated with having all this IoT data in logistics, pointing to the analogy of doctors who don’t want a lot of the data from personal wearables or IoT-generated medical data that could be available to them. Their reason for not wanting all that data is often because of the liability it would bring (“You should have known Mr. Brown was going to have a heart attack without intervention — all the signs were there.”) They asked what responsibility and liability comes with having all that logistics data.
Jim said that Fleet Advantage looks at the performance of the vehicle, via exception reports with parameters that the customer expects their vehicles to operate within to see a potential problem. Sometimes the problem is not the vehicle, but the driver’s behavior. In addition to providing data to the customer, Fleet Advantage also has a Fleet Services team that performs onsite inspections and services on vehicles operating outside acceptable parameters. When asked whether they provide data on what happened when there was an accident, Jim said they have several data feeds from multiple onboard computer and service providers, but are not a black box recorder for accidents.
Ken said “Drivers and their employers are the responsible parties. Trucks with the proper instrumentation can tell you within seconds to minutes that someone is driving in an irresponsible fashion, but I don’t know any case to date where the person collecting that data becomes liable for it. However, that data is subpoenable.” He pointed out that Electronic Logging Devices will be mandatory within two years after the final rule is published by the FMCSA (Federal Motor Carrier Safety Administration), expected the end of this month.1 This will log hours of service and driving behaviors in a ‘black box approach.’ Today the DoT can already force a carrier that is deemed unsafe to install these devices. The current rule does not delve into how the truck is being driven, such as actual speed vs. posted speed, but there is more room for invasive inspection down the road. He thinks the analogy to the doctor is valid — potentially the fleet owner may be liable if they did not take any steps once they knew one of their drivers was driving unsafely.
One panel said liability is an emerging area of concern. More than one panelist said they make predictions based on the data collected and some customers are now asking for money-back guarantees on the accuracy of the data. Another pointed out that data accuracy can be quite challenging for non-IoT data; for example data entered manually by ships’ captains who may fudge their entries to make things look better. Actually that panelist was a little more blunt about it, saying “they lie about everything.” Another said they have implemented algorithms on their inbound data streams to identify suspect data before it corrupts the decision process.
One panelist added that their IoT devices get tampered with in the field. On some of their more theft-prone routes, diversions and theft was affecting 38% of deliveries before they installed their solution. Now with the monitoring and visibility, that rate has been driven down to under 2%. That represents a substantial loss of income for the cargo thieves, as well as for their colluding drivers, so it is not surprising they look for ways to defeat the system. Thus devices have built-in mechanisms to sense and alert to tampering.
Someone asked about the timing for driverless vehicles and whether that would cause a reduction or ultimately elimination of the truck drivers’ jobs. One panelist thought it would take decades before consumer driverless cars were common, but that trucks would happen sooner. Another panelist agreed it would happen in trucks first. Much of the technology is here today. In some cases, such as collision warning and platooning,2 it is available today and part of an incremental adoption path. For example, on long haul routes we may use trucks that do most of the driving for the long distances, while the human driver does the driving on city streets and everything else before getting on the on ramp and after getting off the highway. There are other cases where driverless trucks have been used for years already, such as driverless mining trucks on private property. It was brought up that concerns about the safety of driverless vehicles is largely a perception issue and that driverless vehicles will be safer than human-driven ones.
Will They Replace Truck Drivers?
On the question of whether driverless trucks will mean the end of truck drivers’ jobs, one panelist pointed out that truck drivers don’t just drive the truck. They also load and unload the truck and in any case will act as a backup to the autopilot, similar to pilots in an aircraft. There are a number of truck loading and unloading robots already in use or in design phase.3 Many of these are fairly rigid in the format of cargo they will handle, but just as we’ve adopted to standardized containers, we might adopt much of our cargo to standardized pallet enclosures. Furthermore, the field of robotics is advancing rapidly and robots are becoming increasingly flexible in what they can do.
Nevertheless, some panelist felt driverless vehicles would not fully replace drivers for a very long time. Some thought the technology was not reliable enough yet, but the bigger and longer term issue brought up was liability. The company that is operating a fleet is still responsible if an accident or incident happens and “there’d better be a human driver on the scene to deal with what happens next.“
It was apparent from the discussions that we are witnessing dramatic changes in how our transportation and logistics systems work across all modes — ocean, air, truck, and rail — driven by marrying smart connected vehicles and their sensors with other diverse data sources (weather, social, traffic, news, etc.) and cloud based analytics. The result will be much safer and more efficiently used highways, reduced fuel consumption (and pollution), and much more predictable logistics (and hence leaner and more reliable supply chains). The excitement of these developments was palpable from the panel and the audience and the responses we’ve gotten since the session. This is definitely an area we will have much more to talk about in the future.
_ _ _ _ _ _ _ _ _ _ _ _
1 For more see Final rule for electronic logging devices is pushed back to end of October and Preparing for Complying with the Electronic Logging Device (ELD) Mandate — Return to article text above
2 Platooning is using technology to allow vehicles to safely travel very closely to one another. This increases the capacity of roads, as well as improving fuel efficiency. Trials are being conducted by firms like Peloton and Volvo’s Safe Road Trains for the Environment (Sartre). — Return to article text above
3 Here are examples of truck loading/unloading robots that are already in use or being developed: The Cargo Carousel System, M-900iA 350/T Trailer Loading & Unloading Robot, JOLODA – Trailer Skate Dock, Heavy Duty Robotic Truck Unload System, Paper Reels Automated Trailer Loading — Return to article text above
To view other articles from this issue of the brief, click here.