I collect news daily for my website, The Robot Database. At this time, articles related to autonomous vehicles really dominate the newscape. Aggressive progress is evident related to organization partnerships and acquisitions, state laws to allow the entry of these vehicles, and technology advances related to important component and autonomous intelligence. I like to discuss the future of driverless vehicles with whomever has the interest. The opinions and concerns range as you would expect. Some folks look forward to safer roads, increased convenience, and new opportunities for people who currently don’t have access to automobiles. Other people are less amendable to turning the wheel over to an invisible source that cannot be reasoned with. Obviously autonomous vehicles will be introduced incrementally and deliberately. In this short article I identify some applications that I consider good candidates for early phases of autonomous driving. I avoided precision farming and mining categories as those areas have numerous applications, many of which are already being developed. I will also keep out of the skies. I’ll save autonomous drone applications for another post. With each candidate application in this post, I indicate why I believe it is suitable to the early adoption of autonomous driving.
Occasionally I witness our garbage being collected. The large truck pulls up and a large, hydraulic metal claw grips our hundred gallon container and lifts it until the contents are emptied into the truck. I can see the driver monitoring this maneuver from a small screen on the dashboard. Each time I watch this sequence I wonder how long it will be until the whole operation is automated. I know that companies are already experimenting with driverless garbage trucks, but in the current cases, people are still required to empty the containers into the truck. It’s interesting to note that these trucks are already mechanically equipped with the “robot” required to empty the trash. So the companies that own these vehicles don’t need to throw away that investment. The delta just involves enabling the truck to drive autonomously and automating the task that the drive performs in emptying containers. Given the state of deep learning and object recognition, it’s no stretch to believe that the robot arm within the truck could perform this operation. Challenges would exist for the exception cases of a container that has been knocked over and anything else beyond the “sunny day” scenario. And beyond the technical challenges, it may be politically difficult to replace municipal trash collectors for a variety of reasons. For now, I will consider the case of the privately-run waste management company.
In addition to the technical feasibility, there are some other aspects of trash collection that make it suitable to autonomous vehicles:
- Speed: The vehicles will often travel at a low speed since they make frequent stops. If a municipality was willing to allow autonomous vehicles as long as they maintained speed below a certain threshold, that should not pose a serious limitation to the garbage trucks.
- Time of day: The garbage trucks could travel their routes when the least amount of traffic could be expected on the roads. Traveling in these timeframes would minimize the trucks’ interactions with other vehicles.
- Cost efficiency: Automation is expected to reduce cost in the long term. The cost savings realized by automated garbage collection would be relevant to this service.
- Nature of work: Automation is particularly relevant to jobs that are mundane and dangerous. The job of trash collection is probably more mundane than dangerous but that quality makes the opportunity ripe for automation.
Food and parcel delivery to the elderly and disabled
Not very long ago I was at a community event where the director of the “Meals-On-Wheels” program indicated that their biggest challenge is finding drivers (as opposed to food preparation and other administrative support). I immediately thought: it’s too bad these programs don’t have autonomous vehicles at their disposal. In addition to meal delivery, the elderly and disabled could benefit from cost-effective delivery of groceries and medicine as well.
So this application is suitable to driverless vehicles for several reasons:
- Speed: Similar to the garbage collection scenario, these vehicles would not be significantly limited by speed restrictions; they would not be handling emergency situations.
- Community sentiment: It is likely that a typical hurdle for the introduction of automated vehicles into communities will be overcoming the concerns of the community. But if this particular application serves the community, it may serve as an entry point for driverless cars: the community may decide to allow autonomous vehicles for this specific purpose. And perhaps after a time period of expected minimal incident, the community may be open to expanding the use of autonomous vehicles.
Admittedly there is probably some need for mechanical innovation with the vehicles that handle this task. The would require secured compartments to allow the delivery of the appropriate items to the appropriate recipients. Depending on how many customers a vehicle needed to serve on one route, the compartments may require heating or cooling. And given the nature of the customers being serviced, the vehicles would likely require cameras and some potentially semi-automated alerting system. Finally, this application only involves the delivery of food and parcels. Although there is active research and development related to robots that can provide eldercare, the ability to carry out household work and provide effective companionship does not appear imminent.
Snow removal in office complexes
For now I’m limiting the snow removal job to business parking lots. I believe the job is simpler than handling individual driveways and avoids the likely hazardous public road scenarios. I know that companies such as Kobi are looking into automated snow removal for the private consumer. For this application, I envision a vehicle that may look different from the typical plow-mounted truck. Picture a snow-throwing type unit that throws the snow behind it and caught in an open-top trailer container. Upon filling, the container could be dumped hydraulically (yes, I know first hand that snow is heavy…) on the parking lot perimeter. A complex could potentially utilized a fleet of these vehicles if necessary.
Snow removal offers a good opportunity for driverless vehicles for these reasons:
- Setting: As mentioned above, isolating these vehicles to a private lot could alleviate concern and possibly circumvent state and municipal laws. If these vehicles need to serve multiple corporate premises, they could potentially be moved by person-driven hauler trucks.
- Cost efficiency: Again, we would expect this approach to reduce cost in the long run as automation typically proves to do.
- Nature of work: For obvious reasons, the work of removing snow can be unpleasant, uncomfortable due to the cold conditions, and even dangerous. The timing of the storm may cause the drivers to work through the night. This type of task is suitable to automation due to its nature.
- Leveraging algorithms: I indicated that a corporate site may want to utilize one or more of these vehicles to handle snow removal. Given the number of vehicles involved and the layout of the parking lot, the vehicles could be coordinated by algorithms that assign them the most efficient routes where they avoid each other and dump the snow in a properly distributed manner. The entire fleet solution could sync constantly with weather feeds and human supervisors in case more vehicles need to be awakened or deployed. The assigned routes would be adjusted appropriately. Although a bit complicated, this routing is a task that is solved by a computer much more easily that it could be worked out by human drivers in real time.
Weed harvesting in lakes
I recently vacationed on Lake Minnetonka in Minnesota. While out on the boat one day, I noticed a strange looking barge-like vessel with a large wire basket in front of it. Our hosts explained that the craft was a weed harvester…kind of like a lawnmower for the lake. It was larger than but somewhat similar to the boat pictured here.
The drivers of these boats navigate the lake and remove excess seaweed as necessary. I asked the locals if the seaweed was used to make compost but they didn’t know (or really care all that much).
- Time of day: The harvesting could be done at night when there are fewer boats on the water. The autonomous harvester would rely on GPS and LIDAR for its navigation. I imagine LIDAR and other sensor technologies could be used to find and identify the weeds.
- Nature of work: I have to admit I have never worked as a weed harvester but the task seems straightforward enough to believe that it could be automated. Like the other tasks mentioned the mundaness of the task makes it appropriate for automation.
So there is a short list of applications for driverless vehicles that could lead the widespread presence of autonomous vehicles on the public roads across the country. I understand there are many operational and safety details that I haven’t addressed for each idea. I’ll be keeping an eye out for progress in this area and I’ll be sure to update the list with new ideas that occur to me. Thanks for reading.