February 14, 2020

Aiming to create the world’s first autonomous heavy machinery

Tatsuya Yoshimoto
System Platform Research Laboratories

Yoshimoto joined NEC in 2015 and has been doing research on robotics. In March 2019, his research paper on remotely controlled construction machinery was awarded the 25th Network Research Prize by the IEICE (Institute of Electronics, Information and Communication Engineers) Technical Committee on Information Networks.

Research and development utilizing computing and control

During my time as a student, I conducted research on cyber-physical systems, a transdisciplinary field dealing with computing and control. A big reason why I joined NEC is because I felt like I would be able to utilize the experience and skills I developed as a student. While I was looking for a job, I had the opportunity to meet the person who was the General Manager at the time. They mentioned the company was looking for someone with a talent for programming or control engineering and hearing that became a major reason why I wanted to work for NEC. Another deciding factor was the way the employees felt. I got the impression that a lot of them were serious yet approachable and I felt like they would be a good fit for me.
Even after joining the company, I’ve been able to conduct research which has taken advantage of my specialization as a student. During the first two years of my time at the company, my research was focused on computing. NEC makes great vector processors and I utilized their capabilities to do my research on technologies to accelerate the processing of various software like biometric authentication and social simulation systems.
Starting in my third year, my research theme shifted to the field of robotics and I’ve been conducting research on computer-controlled construction machinery ever since. My goal is to develop an autonomous control system of backhoes by using computers and to fully automate tasks such as the work of excavating dirt and loading.

Successful proof of concept of autonomous construction machinery

Tasks involved in excavating dirt and loading the dirt onto dump trucks, and the operation of construction machinery to flatten surfaces requires the high level of skill and instinct of skilled workers. However, due in part to the effects in recent times of an aging workforce, the construction industry is suffering from a worker shortage. It is much expected that the technology of autonomous operation of construction machinery will help resolve such issues. Additionally, if autonomous operation is realized, there will no longer be a need for people to expose themselves to areas struck by disasters and dangerous construction sites. It will surely help improve safety in the field.
However, the general use of autonomous operation technology of construction machinery in the field is still far from being practical even when one looks at it globally. It’s one of the areas which general contractors and heavy machinery manufacturers are actively working on. It’s amidst such circumstances that we developed our own unique control system called Adaptive Predictive Control Technology and managed to pull off a successful PoC of an autonomous backhoe operation system in 2019. We’re currently making the final adjustments for a PoC in the field.
Incidentally, I look back fondly at the memory of the first time this technology was announced and demonstrated at an exhibition to those outside the company. After all, failure was not an option. We were making adjustments until the very last moment and I’ve never experienced the kind of tension I did when I pressed the start button for the autonomous operation system during the demonstration. The resulting joy and sense of accomplishment that I felt became even greater as a result. After that came the press release and all kinds of newspaper articles. Reading them really made the year and a half which I’d spent working hard on it feel worthwhile.

Establishing your unique place as a researcher

It feels as if the skills required of researchers has been increasing in recent times. In the past, being a researcher was all about honing the core technologies within your area of specialization. However, it’s become important nowadays to not only do that, but to also have the ability to view things from the perspective of a business and to have a solution-focused mindset with an understanding of the needs of one’s customers. Additionally, it’s become essential to broaden your knowledge and to be open to things outside your area of specialization in order to be able to innovate.
However, not everyone can be some kind of superhuman that can do everything. It’s my belief that you first need to be able see how you want to fit in as a researcher and to figure out your place given the way things are. For those of you who are interested in becoming researchers, it would probably be best to start out by forming an image of the kind of researcher you want to be.
As for myself, as one of the few researchers working in the field of control engineering in the company, I’m hoping to demonstrate my potential and establish myself as a unique presence within my field. I’d like to continue to do my best to have a user-focused perspective and to help develop products which will be used by a large number of customers.

A day at work

Message to my past self in my school days