| 研究実績の概要 |
This research proposes the use of "invisible maps" (wireless signal strength and magnetic strength maps) to enable reliable and low-cost localization on existing robots in dense crowds and dynamic environments. A major challenge to making this technology widely adopted is that, due to sensor placement and sensor differences, maps generated by one robot are not an exact match for another.
This year, the effect different devices have on acquired data, maps, and localization has been investigated. For this purpose, several devices, including magnetic sensors, as well as wireless sensors/routers and different antennas, have been purchased. The development of data acquisition units using said devices has been finalized, and have been incorporated into a robot platform for testing. Data acquisition using the robot platform in different environments has begun, with surveyed areas including Utsunomiya University and its surroundings, Tsukuba city, and large commercial complexes in Tokyo and Kanagawa.
Using the acquired data, algorithms to create maps that are independent of sensor characteristics, as well as algorithms to adapt pre-existing maps obtained from different robots, are currently being developed. In addition, a method for building large-scale invisible maps has already been developed and fully tested.
|
| 今後の研究の推進方策 |
Next year will focus on research on algorithms for invariant maps and on expanding testing by using additional robot platforms, which are under development. In addition, it is planed to revisiting last year's environments, so the extent of the effect time has on invisible maps can be elucidated.
|
| 次年度使用額が生じた理由 |
Although research progressed satisfactorily, the COVID-19 pandemic limited access to resources and locations. Therefore experimental testing and some hardware purchases were affected, and international travel for conference presentations was halted. It is planned to compensate for this in the current fiscal year.
|
