The research and development of quantum technologies, which have the potential to solve various present and future social issues, has accelerated both in Japan and other countries. From the perspective of the foundations of industrial competitiveness as well as national security concerns, various national governments are promoting strategic investment in the research and development of such quantum technologies.

In this context, the Toshiba Group has developed and commercialized the following technologies: quantum key distribution based on various quantum technologies including single photon generation, and the Simulated Bifurcation Machine (SBM) capable of solving combinatorial optimization problems as an outcome of its efforts in the research and development of quantum computers. Our approaches to quantum technologies from the fundamental research stage through to social implementation are expected to contribute to the realization of a highly information-oriented, affluent, and safe society.

MURAI Shinya

The rapid progress of quantum computer technologies in recent years has raised expectations for the creation of unprecedented applications, while at the same time increasing the risk of rapid decryption of the encryption algorithms that are widely used in current communication networks. Quantum cryptographic communication technologies are now attracting attention as a solution to this issue, and their social implementation has been accelerating.

The Toshiba Group has been taking the initiative in constructing quantum cryptographic communication technologies based on its long accumulation of development experience and know-how in this field, and is making efforts to promote the commercialization of such technologies. We have been conducting a variety of trial services in various countries toward the inauguration of commercial services, with the objective of realizing quantum key distribution (QKD) services that will make it possible to continuously provide large numbers of customers with secure cryptographic keys into the future.

TOMODA Masanori

In a quantum key distribution (QKD) network, the key management system is responsible for securely sharing cryptographic keys through the use of quantum keys, which are shared among the QKD devices via optical fiber, and providing these cryptographic keys to the encryption devices.

Toshiba Digital Solutions Corporation has been promoting the practical application of key management systems since 2019, in response to market demand for the commercial use of QKD networks. Utilizing the results obtained through various demonstration tests in Japan and other countries, we have been developing vital functions for future key management systems.

Taofiq K. PARAISO / Andrew J. SHIELDS

Quantum key distribution (QKD) exploits the laws of quantum mechanics to offer the highest possible level of communication secrecy. In order to make QKD widely accessible, it is necessary to provide effective solutions to various issues including cost, volume production, and compatibility with standard telecom and datacom infrastructures. Over the past few years, efforts have been made worldwide to provide such solutions using photonic integrated chips. However, the demonstration of a complete chip-based QKD system has remained a considerable challenge.

The Cambridge Research Laboratory of Toshiba Research Europe Limited has achieved a substantial breakthrough by demonstrating the world’s first standalone chip-based QKD system, which is the highest level of system integration of three types of photonic integrated chips. We have confirmed that this system is capable of autonomous operation under practical operating conditions with high stability over several days.

Mirko PITTALUGA / Andrew J. SHIELDS

Quantum key distribution (QKD) is the only technique that allows two users to securely share encryption keys for use in digital information systems with guaranteed security regardless of the resources available to an attacker. QKD relies on the transmission of physical quantum states, or qubits, across an optical channel, often an optical fiber. Unfortunately, the number of qubits that reach the end of an optical fiber decreases exponentially with fiber length. This imposes a limit on the maximum length of current QKD links.

The Cambridge Research Laboratory of Toshiba Research Europe Limited overcame this fundamental limitation by introducing a new QKD protocol in 2018 called Twin Field Quantum Key Distribution (TF-QKD). In 2021, we developed an experimental system and a new phase stabilization technique, called dual-band stabilization, which allowed us to set a new distance record for QKD on fiber, surpassing the 600 km distance for the first time and demonstrating the feasibility of quantum communications over national-scale distances.

Lee JOHNSON / KAWAKURA Yasushi / SATO Hideaki

With the aim of supplying quantum key distribution (QKD) services to the global market by means of QKD networks using QKD systems, the Toshiba Group has been making efforts, in collaboration with partners in various countries and regions, to construct testbed environments in order to implement various demonstration experiments on QKD networks together with user companies.

In the United Kingdom, in cooperation with BT, we have constructed the world’s first quantum-secured commercial metro network and commenced trial services. In the United States, in cooperation with JPMorgan Chase & Co. and Ciena Corporation, we have conducted demonstration experiments on the application of QKD technologies to blockchains in the financial field. In South Korea, in cooperation with KT Corporation, we have constructed South Korea’s longest hybrid QKD network built with different QKD systems over a distance of approximately 490 km from Seoul to Busan, and evaluated its quality of service (QoS) through demonstration experiments.

GOTO Hayato

Various social issues contain combinatorial optimization problems in which the optimal solution needs to be found from among a huge number of alternatives. In this context, attention is being increasingly focused on quantum computers to instantly solve these difficult mathematical problems. Due to limitations on the size of problems that can be solved by existing quantum computers, however, demand has recently been growing for practical solutions making effective use of conventional computers.

During the process of researching its proprietary quantum computer called the quantum bifurcation machine, the Toshiba Group developed the Simulated Bifurcation Machine (SBM) equipped with the Simulated Bifurcation Algorithm (SB Algorithm) as a new algorithm for combinatorial optimization problems. Experiments on an implementation of the SBM on cutting-edge parallel processors taking advantage of its high parallelizability have confirmed that the SBM can solve large-scale combinatorial optimization problems at high speed.

TAKABATAKE Kazuki / KIMURA Keiichi / IWASAKI Motokazu

In the area of drug discovery, quantum computing technologies hold promise as a means of reducing the investment costs and promoting the efficiency of drug discovery processes.

Toshiba Digital Solutions Corporation, in cooperation with Revorf Co., Ltd., has developed a computational drug discovery method to predict the allosteric regulation of proteins with higher accuracy compared with conventional methods. The newly developed method uses SQBM+™, a quantum-inspired optimization solution based on the Simulated Bifurcation Machine (SBM). This method caters to the demand for the application of proteins that have been considered difficult to target in drug discovery up to now, in order to expand the possibilities of pharmaceutical development for difficult-to-treat diseases. As the next step, we are working toward verification of the effectiveness of this method in drug discovery processes through in-vitro experiments.

IZUMI Yasuichiro / MURAYAMA Katsuhito / OKUNO Shun

Toshiba Digital Solutions Corporation has been developing and providing SQBM+™, a quantum-inspired optimization solution that is capable of solving large-scale combinatorial optimization problems with high accuracy and at high speed based on the Simulated Bifurcation Algorithm (SB Algorithm).

We have now launched SQBM+™ Cloud on Azure Quantum, which operates as a cloud service of Microsoft Azure Quantum. SQBM+™ Cloud on Azure Quantum allows users to swiftly realize solutions by using tools and sample programs in the form of software as a service (SaaS), while also eliminating the need to adjust specific parameters of SQBM+™ that tended to place a burden on users.

TATSUMURA Kosuke / HAMAKAWA Yohei / YAMASAKI Masaya

Real-time systems capable of analyzing continuously changing situations and responding to these situations in real time are commonly used in such fields as finance, on-board automotive equipment, and communications. Conventional high-speed real-time systems tend to determine responses in a simple manner due to their time constrains.

The Toshiba Group has developed the Simulated Bifurcation Machine (SBM), a quantum-inspired combinatorial optimization solver that possesses features necessary for real-time systems including deterministic response time and embeddability in edge systems. These features make it possible to realize high-speed real-time systems capable of performing more rational evaluations than current systems, based on combinatorial optimization. We are now conducting various technical demonstrations applying the SBM toward the realization of such systems.

YAMAZAKI Tomohide

The movement toward the realization of virtual power plants (VPPs) that integrate a large number of distributed power sources including renewable energy systems, storage batteries, and electric vehicles (EVs) has recently accelerated. However, in order to match the amount of electricity generated by VPPs to demand, the time required to find the optimal combination of distributed power sources using a conventional computer system often exceeds the practical time limit due to a lack of the necessary computing resources.

To rectify this situation, Toshiba Energy Systems & Solutions Corporation has been implementing proof-of-concept (PoC) studies on application of the Simulated Bifurcation Machine (SBM) to the solution of large-scale combinatorial optimization problems of VPPs. As part of these studies, we have confirmed that the SBM has the capability to swiftly provide better solutions in response to a variety of situations, including steep power fluctuations of renewable energy systems, than conventional computer systems. We have also performed PoC studies on the solution of optimal power flow problems taking the constraints of electric power systems into consideration.

Feature Articles

DOBASHI Hironori

The enhancement of security measures for large numbers of unspecified people in public facilities has become an issue of vital importance. To realize enhanced security in a large space, various problems must be overcome including the difficulty of securing open areas and the need for restraints on the flow of people, which may become a source of stress and inconvenience to them.

The Toshiba Group is engaged in the development of a security system in which a step-by-step screening process is implemented using millimeter-wave radar to ensure safety without restraining the flow of people. We are aiming to offer a security support service utilizing this system to provide the non-contact, high-throughput functionality required for securing large spaces.

AOKI Tomomi / YAMAGUCHI Keiichi / KIMOTO Yoshimasa

Weather radars for the observation of meteorological phenomena including raincloud distribution and rainfall intensity have become indispensable for accurate weather forecasting in recent years. In order to improve the quality of weather radar data obtained at low elevation angles, it is essential to identify and remove ground clutter signals reflected by various objects such as mountains, tall buildings, and so on. Conditional branching algorithms are conventionally used to identify whether received signals are contaminated with ground clutter signals by comparing feature quantities with threshold values. However, engineers skilled in this field are necessary to adjust ground clutter identification parameters at each radar site for the optimal operation of these algorithms. As a result, the need has arisen for reduction of the dependence on skilled engineers and lessening of the heavy burden of such work.

Toshiba Corporation has responded to this situation by developing an automatic ground clutter identification algorithm for weather radars using machine learning, as well as a technique to automatically generate training data for such machine learning from observation data. We have conducted evaluation experiments using the actual data of a multiparameter phased array weather radar (MP-PAWR), and confirmed that these techniques obtain identification accuracies equivalent to or exceeding those of conventional identification techniques.

GANGI Hiro / TAGUCHI Yasunori / INOKUCHI Tomoaki

Low-voltage silicon (Si) power metal-oxide-semiconductor field-effect transistors (MOSFETs) are a key type of semiconductor realizing highly efficient operation of electric power converters and motor drivers. Accompanying the ongoing expansion of the lineup of these products in response to market demand in recent years, it has become necessary to improve the efficiency of design processes in order to reduce the considerable time and effort required for optimizing the design parameters of such a wide variety of products.

Toshiba Corporation has developed a design automation system for low-voltage Si power MOSFETs that is capable of efficiently optimizing design parameters by means of machine learning. From the results of experiments using this system, we have confirmed that it can reduce the actual working hours required for the design of low-voltage Si power MOSFETs by more than 90% compared with conventional manual design methods. Furthermore, as the newly developed system can optimize more than triple the number of design parameters compared with the conventional methods, we have been able to design a new device structure that has 41% lower on-resistance than conventional structures.

OKABE Ryo / YOSHIKAWA Tomohide / SHINOHARA Naoto

In line with the global trend toward the further reduction of carbon dioxide (CO₂) in vehicle exhaust gas emissions, automobile manufacturers have recently been making efforts to promote the development of technologies for 48 V mild hybrid vehicles with appropriate performance to improve fuel consumption while reducing costs.

Toshiba Corporation has developed a 48 V SCiB™ rechargeable lithium-ion battery module with compact, lightweight, and high input-output power characteristics using 5 Ah-class SCiB™ cells as an auxiliary power source for automobiles, and commenced the delivery of trial products to customers. This battery module makes effective use of energy generated by the engine and regenerative energy produced during deceleration, and has sufficient scalability to operate not only as a 48 V battery module but also to realize battery systems of up to the 300 V class for further energy saving through the connection of multiple battery modules in series.

Frontiers of Research & Development

Monitoring Technique to Automatically Detect Deterioration in Accuracy of AI Visual Inspection in Production Lines

*Company, product, and service names appearing in each paper include those that are trademarks or registered trademarks of their respective companies.

Vol. 77, No. 6, November 2022

Index