A wireless robotic system for severe accident applications in Japan

In severe environments, such as a disaster area, various remote controlled robots will need to be used for accurate situational assessment. To manipulate the robots smoothly, video data and robot control data need to be transmitted reliably and efficiently. However, device faults and cable disconnection can easily occur in severe environments.

Communications need to be maintained reliably even under known fault conditions, which may include a physical disconnection, instrument fault or other issues. If a single fault were to occur, the robot is required to maintain communication, detect the fault and, if possible, replace the faulty component.

4.6.2. Wireless system configuration

Figure 15 shows the communication system for remote controlled robots, which was developed as a hybrid system of wired and wireless links. The communication between the robot operation computer and a repeater is provided via a wired network and the communication between the repeater and the robot is done via a wireless network.

Figure 16 shows the abstract structure of the hybrid wired and wireless communication system with single fault tolerance criteria. This system establishes a loop between wired and wireless communication and controls route information by using a rapid spanning tree protocol. Normal communication is established through a wired network. In the case of a cable disconnection, communication is established via the wireless network.

Generally, two network systems are established between the operation terminal and the robot, and these network systems are controlled by using the routing information protocol. In the case of device fault/failure, the system switches

FIG. 14. Maps of the 2.4 GHz signal level transmitted by an access point in HiveManager (left) and FortiPlanner (right).

communication to the other network (the network to which the faulty device is not connected). If wireless interference is encountered, this system has the capability to switch to another frequency. By using these two methods, communication can be maintained even if a cable disconnection or a device fault occurs.

4.6.3. Wireless system prototype

Figure 17 shows overview pictures of the experimental devices. By attaching extension bars, the height of the antenna of the repeater shown in Fig. 17(a) can be adjusted by the robot. For the wireless node on the robot, devices were designed with two different shapes. One is placed vertically (Fig. 17(b)) and the other horizontally (Fig. 17(c)).

The required communication throughput of the robot is up to 20 Mbps. In order to avoid traffic congestion in the 2.4 GHz band, two wireless bands are used for communication. One is a 4.9 GHz band that complies with IEEE 802.11j and the other is a 5 GHz band that complies with IEEE 802.11a.

FIG. 15. System overview of the robot communication system.

FIG. 16. Block diagram of a wired and wireless hybrid robot communication system.

4.6.4. Testing

Tests were conducted based on two types of simulated communication faults: (a) the cable disconnection simulation test and (b) the device fault simulation test. Other tests carried out to check communication performance included throughput and packet loss rate tests. Testing was carried out by simulating the error in the hybrid communication system and observing the performance during a change of robot communication route or system structure as planned. Figure 18 shows the communication error test structure.

4.6.4.1. Cable disconnection simulation test

This test was carried out by simulating the disconnection of cable (2) between the repeaters to confirm the change of the route from wired communication to wireless communication by rapid spanning tree protocol.

4.6.4.2. Device fault simulation test

This test was carried out by simulating the disconnection of cables (1) and (3) of the system alternately to simulate faults in the hubs or the wireless devices, and to confirm the change to the system by the routing information protocol.

The performance of the wireless communication was measured over an area of 40 m × 40 m at an operating NPP. Figure 19 shows the communication performance measurement structure. The communication throughput and the packet loss rate were measured by using the tool ‘Iperf’. The test arrangement for measuring the communication performance was as follows:

— The first repeater was placed at the near point of the equipment hatch.

FIG. 17. Overview of the prototype wireless communication terminal. (a) Repeater (with battery); (b) robot wireless node (vertical);

(c) robot wireless node (horizontal).

— The communication throughput and the packet loss rate were measured when the received signal strength indicator (RSSI) showed −60 dBm, −70 dBm and −80 dBm, respectively.

— The next repeater was placed at a location confirming a throughput of 20 Mbps and a packet loss rate of about 0%, and the above steps were performed again.

— More repeaters were added as above, and the tests were performed again.

FIG. 18. Communication error test structure.

FIG. 19. Communication performance measurement structure.

TABLE 9. RESULTS OF COMMUNICATION FAULT TEST

Test item Result

Cable disconnection simulation test The communication was resumed within 1 s after the cable between the repeaters had been disconnected

Device fault simulation test The communication was resumed within 35 s after the root cable of system A had been removed

The communication was not affected after the root cable of system B had been removed

Table 9 shows the results of the tests. In each test, it was confirmed that the communication was resumed by changing the communication route and the system.

Figure 20 shows the relationship between RSSI and the communication performance in the plant. It was confirmed in the tests that the throughput is over 18.4 Mbps and the packet loss rate is under 2.2% when the RSSI is over −60 dBm.

The communication performance tests in the NPP indicated the highest performance when the RSSI was above −60 dBm. Therefore, repeaters needed to be placed at the points where they would gain an RSSI of above 

−60  dBm.  Repeaters  needed  to  be  positioned  closer  than  this  at  points  where  the  performance  showed  low  throughput even though the RSSI showed was more than −60 dBm. Such points were often found at the junctions  of narrower and broader areas, in corridors with more than two corners and on stairways in enclosed spaces.