Is Ford Putting Its Foot Down on Delivery Robots or Is It Just Pulling Our Leg?

The desire of established companies like Ford, who have been around for over a century, to remind commentators that they are still on the cutting edge plays a large part in announcements like these. However, these automakers—who have proven so successful in scaling up their technology—need to apply their pedigree to robotics.

Digit was spearheaded by Ford Research and Advanced Engineering in partnership with Agility Robotics, a premier developer of bipedal robotic systems. It is made of lightweight material and can lift packages weighing up to 40 pounds, making it suitable for carrying packages to the door. It has a Light Detection and Ranging (LiDAR)/stereo camera navigation solution and can navigate through basic scenarios like walking to a door from the vehicle. When it fails to understand its environment, sensory data is transferred to a cloud platform that allows for the development of further insight. local tele-operation was used to control Digit during the demo, while the picking and climbing tasks were performed autonomously and run locally on the device. The Digit represents a broad aspiration of Ford, to develop innovative last-mile solutions, with driverless vehicles as ad hoc motherships and robots as smaller units in a parcel-to-door solution.

With access to the Ford vehicle and cloud infrastructure to help with compute processing of complex scenarios it is possible to keep Digit lightweight by not overloading it with compute architectures and high-end chipsets. This brings down the price from what is already a highly expensive bipedal robot and extends the battery life. This, alongside the development of AWS Robomaker and moves by Microsoft, represents the continuing emergence of cloud robotics, whereby data is aggregated at the cloud and edge and is increasingly not stored on the device, significantly reducing hardware costs and lowering barriers to entry. However, there remain questions around data privacy and what gets sent to the cloud. There are also questions as to whether a bipedal system is the best form of delivery, given the additional expense and complexity of the machine.

Auto manufacturers are no strangers to extravagant demonstrations of robots. Honda’s ASIMO has gone through a range of iterations, with the latest versions quickly being placed in museums. More recently, Continental, the German automotive parts manufacturer, showcased a similar demonstration to that of Ford and Agility Robotics except it had a quadruped from ANYBOTICS. ANYBOTICS is in many ways the European equivalent to Boston Dynamics, which has been keen to emphasize its utility for last-mile delivery purposes as well as for construction. In mid-2018, Boston Dynamics indicated that it hoped to build over 1,000 SpotMini robots over 2019. As of ProMat 2019, its main demonstration was a different robot for material handling for warehouses as well as a nonmobile solution. At present, the SpotMini and its deployment across construction sites has been little more than marketing.

There is no denying the impressiveness of the technology being showcased. However, given the myriad failures taking place in the consumer robotics industry with the fall of Anki, Mayfield, and Jibo in the last year and a half, the best way to reassure stakeholders is to focus on practical solutions that are ready for the market now.

Robotics is not a novel technology anymore. Given the saturation of technology demonstrations (a prime example being the walking car hybrid showcased by Hyundai at CES 2019), these stories should be replaced by actual testing grounds on campuses, in workplaces, and in specific neighborhoods before the market can take more notice of them.

In the case of Digit, there are certainly advantages to bipedal systems. They are more agile than quadrupedal systems and have less trouble with uneven terrain then standard four-wheeled Autonomous Mobile Robots (AMRs). However, they also suffer from lack of stability, lower payload, and mechanical complexity. Also, AMRs are generally far more expensive to deploy; hence there is a lack of commercialization.

Automotive OEMs like Ford, GM, and others are engaged in a far more extensive and strategic technology shift than these relatively limited robot-related Proof of Concepts (PoCs). The push is not only toward autonomous vehicles and haulage but also toward smart mobility services. In this sense, the rather “exotic” nature of the bipedal robot is not as important as Ford’s desire to build autonomous fleets that will fulfill last-mile delivery purposes.

Ford was among the first mass market OEMs to make the case for a transition to smart mobility services. In 2016, Ford decided to form a new specialist division called Ford Smart Mobility to develop and deploy new shared mobility services as well as make strategic investments to grow Ford’s smart mobility interest. In that same year, Ford acquired Chariot, a public transit-style carpooling service that operates shuttle services along set routes using Ford vans. This was followed by the acquisition of Autonomic, a connected car sensor ingestion platform designed to absorb, normalize, and enrich data from connected cars to facilitate fleet management, maintenance, and multimodal transport. Ford also acquired TransLoc, a startup focused on providing demand response software and simulation tools to transit operators—confirming Ford’s interest in offering a transit-like service and in offering future drivers Containers-as-a-Service (CaaS) solutions within existing transit options.

Though behind GM and the tech giants on autonomous vehicle development, Ford has also established a number of strategic partnerships to support its autonomy objectives. In late 2017, Ford announced a partnership with Lyft to deploy Ford’s driverless vehicles on Lyft’s ride-hailing platform; deployment is targeted for 2021. Ford added to these partnerships at CES 2018 when it outlined a smart city-centric vision, detailing partnerships with Domino’s and Postmates to explore alternative use cases (in this instance, food and goods delivery) for its vehicles. Domino’s has also partnered with Starship Technologies, one of the more advanced developers of autonomous robots for last-mile delivery. Starship currently has an app available and is operating in a select number of locations, including Milton Keynes (UK); Tallinn (Estonia); and George Mason University, Arizona University, Napa Valley, and Oyster Bay (United States). Ford should prioritize partnerships with more standard robotic form factors that already have products ready and play the long game with bipedal and quadrupedal systems.

Drones also represent an interesting opportunity for smart mobility, as showcased by PoCs that had drones launch from vehicles. But these examples are hampered by regulatory hurdles, the likelihood of crashes, and questions of payload. ABI Research values the delivery Small Unmanned Aircraft System (sUAS) opportunity at over US$10 billion by 2030, as noted in recent ABI Research sUAS Market Data forecast The Small Unmanned Aerial System Ecosystem (MD-SUAS-105). ABI Research expects a smaller revenue for ground-based delivery for nonautomotive robots at this point, with US$9 million in actual revenue coming from robotic delivery applications globally in 2019. This does not cover intralogistics or delivery within hospitals or office spaces; it only covers consumers.

In essence, Ford has already developed a lot of the pillars necessary for smart mobility services, which will become increasingly vital to their growth and competitiveness in the future. The company has identified last-mile logistics as a huge market worth hundreds of billions of U.S. dollars in future revenue. Ford has stated that it is not planning to compete with FedEx or UPS on commercial delivery and that the Agility partnership remains in piloting for the foreseeable future. Ford could augment this partnership by testing it on ever larger scales in the same locations, perhaps in concert with a more advanced partner like Starship Technologies.

In addition, Ford might consider mixed fleets of robots as a more comprehensive solution, in which case its development of advanced fleet management and back-end data analytics competencies could be invaluable. A mixed fleet last-mile delivery solution with robots and vehicles—even on a small scale—not only would require the assistance of robot manufacturers but also could include robotic software providers and autonomy solution providers like Brain Corp, Raven Ops, and Freedom Robotics.