Qi Kong is Chief Scientist and Head of Autonomous Driving at JD Logistics, a business group under JD.com, China's largest retailer.
After the sudden outbreak of COVID-19, in order to remain safe and healthy, people had to stay home and avoid unnecessary human interactions, which led them to rely heavily on online shopping.
On February 6, JD.com fulfilled its first order with an autonomous delivery robot in Wuhan, the epicenter of the COVID-19 outbreak in China. The robot demonstrated its value by delivering daily necessities and medical supplies to hospitals and residential compounds for both medical personnel and customers. JD also leveraged drones for delivery to areas where traditional delivery routes were temporarily suspended.
As even online shopping has traditionally involved elements of human interaction, such as parcel delivery, companies have lately been using contactless delivery methods, such as code-controlled lockers, mini-stations outside of residential compounds to pick up and send packages and especially delivery through drones or autonomous robots.
These autonomous robots are not a new technology but are still in a relatively early stage, and we have a way to go to achieve operation at scale. What are the challenges that hinder the fast deployment of autonomous robot delivery? And are they cost effective compared with human labor? As long as we address three issues, the widespread use of autonomous delivery robots is well within our reach.
First, it takes time for the autonomous technology to mature. Many of the companies which are developing robot delivery mainly test their robots in closed parks, where the roads are simpler and safer, rather than on public roads.
For a public road with a more complicated traffic scenario, the technology must be similar to that of a robotaxi, which uses highly autonomous driving technology based on or able to simulate real road conditions. JD has pioneered combining this technology with real logistics scenarios.
Robotaxi technology requires a higher level of safety verification as the taxi needs to worry about the passengers in the car as well as pedestrians. Using a cloud-based simulation platform, engineers can optimize the technology and test the validity of a feature in just one hour. This greatly improves the efficiency of technology upgrades. Previously, verification through real road scenarios would have taken a year.
With further development, the robot can adjust its speed according to real road conditions, for example slowing down when it encounters a passenger or speeding up on open roads, ensuring smooth autonomous driving. It becomes likes an experienced driver who can adjust to more complicated scenarios, but it needs time and scale to work.
This is one of the reasons JD was able deploy robot delivery in Wuhan with just five days of preparation, as the research and development for this technology was handled remotely by JD's team in Beijing.
The second issue is that companies need to identify scenarios which might require autonomous delivery robots. While COVID-19 is by no means welcomed, the pandemic and the urgent need to respond to it have created an opportunity for autonomous delivery robots to demonstrate their value to society.
Because COVID-19 is highly contagious, any human interaction is considered risky. As the country with the highest e-commerce penetration rate in the world, Chinese consumers rely heavily on e-commerce platforms to buy almost all their daily necessities, putting pressure on logistics, especially during the Chinese New Year holiday when only a few companies continued normal operations.
A JD.com delivery staff sorts out parcels outside a residential compound in Beijing on Feb 21: Chinese consumers rely on e-commerce platforms to buy their daily necessities, putting pressure on logistics.
© Reuters
The deployment of autonomous delivery robots can not only protect couriers and consumers but also help reduce couriers' workload. On its arrival, the robot will send a message to let the customer know, and they can pick up the parcel by entering a code on the robot's screen or by using facial recognition.
As the problem of aging populations gets worse in some countries, and labor costs increase around the world, autonomous delivery robots may help to solve some of the labor shortages.
In this case, the robots can perform simple and repetitive deliveries and other tasks, freeing up humans to do more complicated deliveries, such as a bulky item to a customer's doorstep. A robot can help send a batch of items to an office building and wait for customers to pick up the orders at a fixed point, or work as a mobile warehouse for couriers and wait where they want.
Robots can also be deployed in other scenarios, to serve as a smart restaurant waiter, coffee delivery machine, fire warning device and more.
Finally, regulations concerning the use of these devices have yet to be finalized and announced. Safety is always the top priority, which is why, before deploying autonomous delivery robots on the road, approval from local authorities is needed.
The trend is that as stakeholders adopt an open attitude toward these regulations, more and more companies around the world are allowed to test autonomous delivery robots on the road. The clearer the policies, the more possible it is to quickly scale operations.
Autonomous delivery robots, which provide a contactless service that protects couriers and customers, have made customers more comfortable with last mile delivery during COVID-19. Following the successful experience with our first robot in Wuhan, we quickly deployed a second and will consider adding more to the fleet nationwide as appropriate.
I believe that this is the era for accelerating the deployment of autonomous driving at scale.
