Cyber Physical System
Laboratory
We aim to realize a data-driven society data-driven society
that integrates cyber space and physical space
by utilizing sensing technology, data analysis and machine learning.
Faculty of Software and Information Science, Iwate Prefectural University
Horikawa Lab
Topics & News
2023/02/16 Presentation at the New technology briefing session (YouTube)
2022/11/07 Student's Project on the our university's official channel Video (YouTube)
2022/10/18 CEATEC2022 Exhibit contents (PDF) / demo video (YouTube)
Our Research
01
Smart factory
We are developing technology to visualize the entire factory using sensing and machine learning. Currently, we are mainly working on worker behavior analysis.
We have developed technologies for "estimating worker behavior in a fixed position" using video analysis and "analyzing the behavior of workers on the move" using small sensing devices, and have been involved in demonstration experiments.
As a result, the small sensing device is now being sold as "InQross Kaizen Maker" by our joint research partner, ERI Inc. Please refer to Youtube for these demonstration videos.
Currently, to solve the issues in the demonstration experiment, we are working on the following three issues by using machine learning in combination with video analysis and sensing devices.
(1) Motion estimation using unclear videos
(2) Automatic labeling of correct answers to videos using sensing devices
(3) Individual identification of workers with similar appearances
03
People flow analysis
At the request of a company, we are working on people flow analysis using Wi-Fi probe requests.
People flow analysis using Wi-Fi probe requests has been proposed in the past, but it has become difficult to use due to the randomization of Mac addresses to enhance the privacy of mobile terminals.
In response to this, we are developing a technology that uses machine learning to estimate the number of terminals in an area using only the information in the Wi-Fi probe request without using the Mac address.
Currently, we are conducting a demonstration experiment assuming a dense smoke space in the event of a tunnel fire, but since it is a technology that can be measured with a simple computer such as RasPI, we plan to consider applying it to marketing, such as grasping the number of people staying at the store. is.
05
Digitalization of
local activities
This is a project-type research to be undertaken as a regional collaborative research (2023-2024) with Yamada Town, Iwate Prefecture.
It will be a project that utilizes digital technology for community activities while considering the role of community associations (neighborhood associations), which is an issue nationwide.
Focusing on third-year students who have just been assigned to this laboratory, we will work on the following three things.
(1) Utilization of ICT to support community association activities
(2) Utilization of SNS for visualization of local activities
(3) Human resource development for the spread of digitalization
02
Webpersonalization
Web personalization technology, which delivers information suitable for users based on their browsing history, location, and time, is widely used. In this research, we are trying to create new web services by utilizing "user status" for web personalization.
We have established the following two technologies using acceleration/gyro sensor data obtained by JavaScript in web browsers.
1. User motion estimation (walking, stairs, lying down, standing)
2. State interval estimation (apply change point detection method)
Currently, we are creating a prototype that combines the above two into a "state" and aims to dynamically generate an interface according to the state, recommend EC sites, and increase the average stay time.
Utilizing this technology, students in the laboratory developed a health promotion web application in collaboration with local companies.Click here (Youtube)Please.
04
Utilization of sensor data
We are promoting research that uses various sensing data in machine learning.
Case 1) Improvement of indoor positioning accuracy by multimodal learning
We are aiming for a positioning error of less than 1m with beacon positioning using radio wave intensity. Currently, we are working on the development of a high-precision positioning method by establishing an automatic generation method for training data and multimodal learning of radio wave intensity and acceleration/angular velocity data.
Case 2) Worker behavior prediction using sensor data and GNN
I am researching the behavior prediction of factory workers from sensor data and graph neural networks. By predicting short-term bottleneck processes, we are considering applying it to next-generation production management systems.
Case 3) Visualization of the relationship between human behavior and communication
We are conducting research to visualize the relationship between human behavior in the office and the resulting amount of conversation.
06
Transmission using
cosmic rays
We are helping out with the Mt. Iwate Penetration Project using cosmic ray muons.
This research will use machine learning to analyze whether the bedrock density data of Mt. Iwate collected from the muon observatory can be used for eruption prediction and observation of volcanic activity from time-series analysis.
Past Research
07
Robot arm
We have developed and commercialized a "robot arm type application operation monitoring service" for DNP Communication Design, in which a robot arm monitors the operation of smartphone applications 24 hours a day, 365 days a year on behalf of a person.
This is an example of how we applied the results of our research on image recognition technology and RPA to a corporate problem.
08
Drone
To improve the accuracy of unmanned drone landings, we have developed a radio guidance system for drones using beacons.
We are developing automatic tracking navigation and navigation via relay points.
I think this technology will be necessary when logistics delivery by drones becomes widespread.
09
Beacon positioning
This project is the origin of the industry-university collaboration in our laboratory.
Developed with DNP Digital Com (now DNP Communication Design), it was used as a positioning engine in indoor navigation demonstration tests at Tokyo Station and Shinjuku Station.
This technology is still being used in smart tags and other applications.
Education
Research policy
We are conducting research to develop technologies that are useful in the real world.
While monitoring technology trends at top conferences and other events, we consider the elemental technologies that will be needed in society several years from now, and while promoting our own technology development, we present our findings at conferences in Japan and abroad, submit papers, and apply for patents.
At the same time, we actively support joint research and commercialization to solve social issues using elemental technologies by monitoring social trends and receiving proposals from companies.
We are working on our research with the spirit of, "Conferences are just a passing phase, let's innovate to change society! This is the spirit in which we conduct our research.
Educational policy
We proactively incorporate design thinking and agile development, and make "How to create innovative services? as the pillar of our education.
Our weekly seminars are more of a place for members to share their knowledge than a class, with planning meetings and sharing of the latest topics.
In addition to research activities, we actively encourage members to participate in voluntary activities (external contests, long-term internships, PBL activities, etc.).
Technology
We encourage student-led patent applications.
Since 2014, there are 7 pending applications and 5 registered.
Most of them are independent applications, and after filing a patent application, we often provide the technology to companies for use.
Contact
For access to the Takizawa Campus, please see here.
*If you come by car, you can freely use the parking lot.
Near the B building of the Faculty of Software and Information ScienceIt is convenient to stop here.
You do not need to go to the reception desk to enter the campus.
Faculty of Software and Information Science, Building B, 4th Floor Laboratory 20 (Horikawa Lab) (see map below).
152-52 Suko, Takizawa City, Iwate Prefecture 020-0693
Faculty of Software and Information Science B Building 4F Research Room 20
Tel: 019-694-2500
Fax: 019-694-2501
