HTML5 Webook

70/182

damaged. It also should be noted that such connections are not available free of communication charge.With regard to method (2), systems using satellites such as IMMARSAT satellites have been available in the market; so, although limited to open air areas where satellite signal is available, we are able to conduct drone operations glob-ally not constrained by locational conditions—on the sea surface or on a mountain. However, the method has major drawbacks: its operational cost will be a big matter to insuf-ciently funded users, and in addition, drones have to be large enough to have such satellite communication systems mounted.Method (3) is expected to oer lower cost than methods (2) and (3). Although the conventional technologies for controlling robots by wireless relay are available, they are based on the wireless LAN technologies developed for the Internet; so, at each timing of relay path switching the control communication is interrupted for several 100 ms to over 1 second, which leads to the problem that the robot will be uncontrollable during such a period of communica-tion cut-o. Another problem is that there exists a time delay uctuation from command transmission to command reception (by a robot), and the control response delay time (latency) is not guaranteed. Consequently, such systems are not fully applicable to robot control.We have promoted our research and development ac-tivities on the assumption that our system will be used in a situation as shown in Fig.11 where a robot beyond line-of-sight is to be controlled and monitored through other robots acting as relay stations. Our system should be dedicated to “robot control,” and our system uses “relayed communication,” to design and develop an advanced access protocol that ensures keeping the latency under a certain value and avoiding communication interference [16].We successfully guaranteed the latency by applying “Time Division Multiplex Access (TDMA)” to robot con-trol, where to each of the communication paths—control station to relay station, relay station to relay station, and relay station to robot—a pre-dened timeslot is allocated. Furthermore, we applied the following scheme to robot control for the rst time: instead of using the conven-tional procedures to nd or determine communication paths prior to conduction of communications, every in-coming signal to each slot is always received, and then sorted and accepted according to a rule where the strongest signal is accepted, or another rule where signal acceptance is determined by a predened priority.rough applying these technologies, we accomplished the following: suppressing the latency in a case of relaying though a relay station, which shows dierent values de-pending on operation conditions when a conventional method is used, within a control signal transmission pe-riod (about 60 ms), and enabling avoidance of instability in control; suppressing the duration of communication cut-o occurring when the relay path is altered as the robot moves, less than a tenth of the cut-o duration by the conventional method, which means that we enabled keep-ing the “freshness” of the control data received by the robot within a certain level even in a case where the control signal is relayed by a relay station on the control data path. We have been calling these technologies “tough wireless” because they are resistant to radio wave blocking and us-able in a disaster situation.e wireless devices we developed use frequencies or channels as follows: using a 920 MHz-band small power radio station (compliant to ARIB STD T108) for both control signals and telemetry signals bidirectionally; bun-dling up to ve 200 KHz width channels within a width allowed by the standard and using the bands as a 1 MHz width band; improving the interference resistance by se-quentially switching four frequencies one to another. In the proof-of-concept experiment eld tests conducted in June, November 2016 and June 2017, we successfully conducted stable operations and telemetry signal reception of a small 4WD robot or a multi-rotor type drone existing in a loca-tion of beyond line-of-sight from an operator and at the same time beyond radio line-of-sight. During the experi-ment, another drone with a relay station mounted was hovering about 20 to 30 m above the ground. We conrmed that we were able to control the target drone, which had FiF11Beyond line of sight operation through multi-hop relay by drones/robotsRelay robot(in flight)Relay robot(on the ground)Target robot(in flight)Target robot(on the ground)Remotecontrol operatorObstacles including concrete buildings, walls, woods, or mountainsObstacles shielding or attenuating radio wavesInvestigating buildings or the areas on the other side of a mountain in disastersDelivering goods or medicine/medical supplies to villages in mountains2 Terrestrial Communication Technology Research and Development66　　　Journal of the National Institute of Information and Communications Technology Vol. 64 No. 2 (2017)