Object recognition, localization and private object interaction tracking for maintenance-free IoT
| Project/Area Number |
22K17883
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 60060:Information network-related
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
エルデーイ ヴィクトル 大阪大学, 大学院情報科学研究科, 特任助教(常勤) (40850938)
|
|---|
| Project Period (FY) |
2022-04-01 – 2026-03-31
|
| Project Status |
Granted (Fiscal Year 2022)
|
| Budget Amount *help |
¥4,680,000 (Direct Cost: ¥3,600,000、Indirect Cost: ¥1,080,000)
Fiscal Year 2025: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2024: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2023: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
Fiscal Year 2022: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
|
|---|
| Keywords | Object recognition / wireless sensing / localization / backscatter tags / object recognition / privacy / interaction tracking |
|---|
| Outline of Research at the Start |
We aim to develop technological solutions to improve our awareness of our environment, with a focus on our interactions with objects. Such information has several applications including healthcare and well-being (lifestyle analysis, navigation assistance), asset tracking, and security screening. These applications require us to recognize, identify and track objects and their interactions. In this project, we will:
(1) Build a contact-less object recognition system for plain, uninstrumented objects
(2) Develop a scalable, low-power, privacy-preserving system for tracking object interactions
|
| Outline of Annual Research Achievements |
Object recognition: We prepared and submitted a full paper to the IEEE IoT Journal. We developed RadioRec, a deep learning-based system that recognizes everyday objects based on their interactions with microwave signals in a contact-less manner. RadioRec is a significant upgrade compared to our earlier workshop paper, both in terms of the learning capabilities and in terms of evaluation. RadioRec works by transmitting a microwave signal through the object using a single antenna pair. RadioRec extracts features automatically using an autoencoder, and uses them to train an object classification model. Our evaluation shows that RadioRec can detect and recognize 26 everyday objects of various materials and shapes with an accuracy of over 97%.
The paper was rejected. The reviewers requested additional evaluation in a different environment, and also several clarifications in the text. We are now working on addressing the reviewers’ comments and preparing an updated submission to the same journal, as suggested by the reviewers. So far, we have performed additional experiments in another environment, with various object orientations and object positions, and have obtained promising preliminary results.
Object localization: We have prepared a preliminary implementation of SpotFi (DOI: 10.1145/2785956.2787487) for the USRP software-defined radio platform to enable the simultaneous estimation of angle of arrival and time of flight on backscatter tags. We are currently working on getting the estimation to work correctly in a controlled environment (for now, without backscatter tags).
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
Based on the details reported in the achievements section, we can say that the research itself has progressed more or less according to plan. However, the publication of our object recognition paper is being delayed due to our initial submission having been rejected.
|
| Strategy for Future Research Activity |
Regarding object recognition, we will continue to address the reviewers’ comments related to our rejected journal paper and work on resubmitting the paper.
Regarding localization, we plan to continue working on the USRP-based implementation of the SpotFi algorithm. We will try to achieve successful estimation of angle of arrival, and time of flight, first in very controlled environments, then gradually moving towards less controlled environments (e.g., ones that have multipath effects). Then, we plan to use the obtained estimates to determine the location of the target.
Wa also hope to explore the idea of enabling privacy-enhanced encounter-based interaction tracking with backscatter tags as outlined in the original research proposal.
However, we have also identified new application possibilities related to object recognition and localization. In the field of wood processing, there seems to be a need to identify the presence or absence of knots and their 3D location in unprocessed logs. With this information, woodworkers can determine the optimal cutting method. Knowing the location of knots inside the log, which is difficult to determine from the exterior, allows the sawyer to avoid knots and produce high-quality knot-free boards, increasing the value of the wood while reducing waste. We may experiment with various types of electromagnetic waves and wireless imaging techniques to perform this task.
Since this application has significant potential for direct use in society, we might prioritize this direction in our future work (depending on discussions with stakeholders).
|
Report
(1 results)
Research Products
(1 results)
