Development of an actuation system based on water jet propulsion for a slim long-reach robot
© The Author(s) 2017
Received: 5 November 2016
Accepted: 22 February 2017
Published: 7 March 2017
Most of the long-reach inspection devices developed so far have a limitation in the reduction of their diameter when their length is increased. This limitation is due to the technology used to provide the actuation to the system. The use of water-jet as propulsion source is promising to be a solution for this problem. Currently different devices use water-jet as propulsion. However, none of these systems have been designed for small-scale nor have a straightforward control. Therefore, in this paper, we discuss the development of a long-reach water-jet probe aimed to be used in Fukushima Daiichi Nuclear Power Plant. The main elements that define this probe are three high-pressure pumps at the base, three hoses whose water flow is controlled for maneuvering the device, an Inertial Measurement Unit to acquire the attitude of the tip, and a joystick that allows the user the control of the whole device. Moreover, its design allows it to move in different kind of environments and generate three-dimensional motion. Besides the experimentation developed to characterize its behavior, the system was tested in different environments; such as on the ground, in the air and inside the water; showing the best control in aquatic environments.
KeywordsInspection robot Long-reach robot Water-jet propulsion
The incident at the Fukushima Daiichi Nuclear Power Plants (NPPs) has brought different kinds of challenges to the development of robotic devices that are used for decommissioning and cleanup of the installations. Several robots have been used for the decommissioning of reactors in places where accidents have happened, such those employed at Three-Mile Island , or for the dismantlement of different facilities around the world due to the end of their lifespan [2, 3], or just for quotidian maintenance labors . All these robots have in common that most of them have been designed to perform certain tasks in environments that are well known or require the preparation of the surroundings to perform their task without problem. Unfortunately, due to the way the accident in Fukushima Daiichi Nuclear Power Plant occurred; earthquake, tsunami, and the hydrogen explosion of the reactor buildings; the infrastructure of the nuclear facility was severely damaged, making it extremely difficult to access critical areas. Ascertaining the exact condition of theses areas has become very complicated. Therefore most of the robots developed up to that moment were not able to be deployed safely inside the facilities.
Debris blocked the access to several areas and many places were flooded due to the tsunami. Various critical locations inside the NPP such as the reactors were inaccessible and their statuses were unknown since many of the measurement instruments were also damaged. This brought severe problems since it was extremely difficult to know the correct way to proceed for the decommissioning and the solution to other important issues, such as stopping the leakage of radioactive material. Before the incident, there were not robots specially designed to deal with this kind of tasks or conditions. For that reason, some of the first robots used to survey the disaster zone were originally designed to be used on military operations . Although these robots were adapted to be deployed in the disaster area, they faced different difficulties or limitations and the design of more specialized robots became necessary .
The robot deployed in unit 1 is slim enough to access through the piping. It uses crawlers as propulsion elements. Furthermore, as is indicated in Fig. 3b, it is equipped with sensors such as a thermometer, dosimeter, CCD cameras, and laser for measuring distances and is remotely operated by a wired cable. As is shown in Fig. 3c, it is designed to change its form in order to generate better mobility when it reaches the grating. In order to verify the location of fuel debris bellow the grating, it is important the development of a device that can access through the grating whose shortest spacing length is 25 mm . Furthermore, since the current robots are propelled only through crawlers, which require a continuous contact with some surface to properly work, their mobility might be limited in the water contained under the grating. Taking into consideration the condition mentioned before and the limitations of the systems used so far, it is required an inspection system slim and long enough to gain access through the piping and overcome the grating. Moreover, this system must be tough enough to bear radioactivity and aquatic environments, as well as good controllability under water.
Among the different robotic systems available at the moment of the accident, those known as continuum robots seemed to be the best option for this task since these systems are highly adaptable to the surroundings and possess a high length/diameter ratio compare to other systems. Furthermore, in most continuum systems, the heavy actuation elements are placed at their base, enabling the isolation of these components from the harsh conditions of the environment where the inspection is performed and reduction of the weight of the device’s inspection segments. In general, these kinds of robots are classified as intrinsic, extrinsic, or hybrid . These classifications are based on the kind of actuation they use. The intrinsic configuration uses fluid in bellows, which allows expansion or contraction by the change of pressure in the fluid. In this manner, the bellow defines the body and actuator element for this system. By the arrangement of at least three bellows in a parallel array, it is possible to generate spatial motion. Additionally, by serially connecting these arrangements, it is possible to build a longer device with more complex locomotion capabilities.
Some examples of this configuration are the systems developed by Festo  and the OctArm . On the other hand, the extrinsic configuration is based on bending flexible backbones, such as springs or rods, that are connected at the end of the elements through tendons. These tendons come in triads for every elastic element in order to generate spatial motion. As with the intrinsic systems, several groups of these elements can be connected serially to increase the length and complexity of locomotion. Some examples are the robots developed by OC Robotics  and the tendril robot developed by NASA . Finally, the hybrid configuration uses a mix of the systems mentioned before. These combinations could be a concentric/pneumatic, where a flexible backbone is bent through pneumatic actuators instead of tendons; or tendon/pneumatic, where the backbone is based on pneumatic actuators and tendons are used to bend the body .
While continuum robots generally have good length/diameter ratios, an increase in the robot’s length typically necessitates an increase in the diameter of its proximal part. This is because with every new segment added, more tendons or hoses are required. All these elements have to pass through the body from the base of the system to the segment where are acting. Therefore, the segments nearer to the base tend to be thicker than the rest of the body. This is a disadvantage considering that this task requires a considerably long device that should move inside a 100 mm diameter piping.
OC Robotics has developed a series of snake-arm robots that have been used successfully in NPP to conduct inspection or repair operations . Despite the fact that these robots have high-performance mobility and some of them have diameters smaller than 100 mm, their length is limited to a couple of meters. Another project under development, which also uses wire-rope drive, attempts to increase reach length of the arm. However, the proposed final diameter is not small enough to be deployed through the piping . From the examples mentioned above, the tendril developed by NASA is the only system that keeps a slim constant diameter throughout its length . This is because only a few sections at the distal part are actuated and the rest of the body is passive. This design limits the locomotion of the system and it does not allow controlled propulsion nor locomotion in open areas since the only propulsion force forwards comes from pushing the tendril from its base, where the system is flexible and lacks actuation.
A robotic system that allows the inspection of narrow spaces while keeping a constant diameter all along the device is the active scope camera . This robot has been tested in disaster situations with debris and has shown good performance in adapting to its surroundings. Its propulsion is based on the vibration of its cilia. However, it requires reasonably good contact with some surface in order to have good propulsion. Taking into account this point, this system would have a limited control or would lose its propulsion due to the lack of contact with surfaces after it exits the access pipe and is inside the water.
Other long-reach systems with a small diameter are those used for the inspection and maintenance of piping networks. The body of these systems is essentially a passive elastic hose, flexible enough to adapt to the continuous change of direction present in pipings, and its propulsion source is located at the tip. This propulsion is based on water-jet which generates enough thrusting force to pull the whole system inside the network. The movement of these devices is limited since their propulsion is only forward; some special passive tools are used to assist in adapting to changing directions . In case the pipe is big enough and path selection is necessary, robotic mechanisms can be employed . Despite the fact that these systems have a fixed diameter along their body and capable of self-propulsion. They lack control for open areas, which is necessary when the system is inside the PCV. There exist other devices that use water-jet as their propulsion source and it is possible to control them in open areas. These devices are developed for amusement activities. Their size is considerably large and their control is completely manual. Furthermore, the user requires highly trained skills, as well as ride the device to control it through his/her body movements [19, 20].
High-pressure pump. This element is necessary to generate flow through a long and thin hose. By the regulation of the speed of the pump, the flow is regulated, consequently the thrust force as well.
Flexible hose. This element allows the transportation of water from the base to the tip of the system. It should be strong enough to bear the pressure required to transport the liquid, as well as flexible enough to allow adaptation to the surroundings. Furthermore, it should be long enough to move through the 20 m length pipe and conduct the inspection inside the PCV.
Nozzle. This element is used for the redirection of the flow, as well as to increase the speed of the fluid in order to augment the thrust force of the water-jet.
Posture sensor. This sensor is to acquire the variations on the attitude at the tip of the device. Data obtained from this sensor is used to adjust the flow of water in order to correct any variation that modifies the operator’s commands. Moreover, we consider the addition of more nozzles along the body. These will assist the movement of the system, in case an inspection in a more complex piping is planned. Other sensors such as camera, dosimeter, temperature sensor, and so on, should be mounted at the tip of the probe in order to have a useful inspection system.
In this paper, we center our attention on the study of the feasibility of the water-jet as a propulsion for a slim, long-length inspection system and the analysis of a device potentially capable of fulfilling the requirements of size and mobility necessary for the inspection task inside the PCV. In this first stage of the study, we focused on the propulsion and control at the tip of the device. First, we outline different possible configurations of devices that allow us to verify our proposal and then the theoretical background is formulated. Next, we analyze the results of basic experiments performed in order to understand the relationship between the different variables and how they affect the propulsive performance of the system. Afterward, considering the results of the characterization of the variables, a first prototype is suggested and the details of its mechanical, electronic and control design are explained. Finally, a discussion and analysis of experiments carried out in order to verify the mobility of the prototype in different environments is presented.
We consider the water-jet as a good option to be implemented as propulsion source for a probe since a water-jet can be generated via the transfer of water through a long hose with constant and relatively small diameter from a water source to an output nozzle. Additionally, such a water-jet generates a thrust force that can be controlled by the regulation of the flow rate. Taking into account that this system requires control in open areas, three different configurations are suggested. The first one is a configuration similar to the tendril from NASA with the difference that only one actuated section is required at the tip; in this way, it is possible to position the tip in the desired place in the three-dimensional space. The water-jet is directed partially backwards to the longitudinal axis of the hose, in the same way as in the piping inspection devices, and via the regulation of the flow rate, the magnitude of the thrust force can be controlled. The second option is by regulating the water-jet through servo-valves. Similar to continuum robots, at least three actuation elements are necessary to control the motion in three-dimensional space. Thus, it requires independent control of each water-jet. This necessitates at least three servo valves which should be placed at the tip of the probe in order to regulate the opening of the orifices of each water-jet, thereby controlling the flow rate and thrust force. However, the need of servo valves at the tip will increase the diameter of the device and the complexity of the manufacture. Finally, the last proposal is a device with no mobile element at the tip. Instead, it has three nozzles in circular arrangement directed partially backward. Each water-jet is generated independently, which means that each nozzle has its own feeding hose and independent pump. By regulating the speed of each pump, the flow and thrust force are controlled. Because the last option is the easiest to implement, we chose the three hose configuration to test the concept. A more detailed study of the generation of water-jet thrust force is presented next.
The understanding of the thrust force generated at the tip of the system is crucial because this force is used for the forward propulsion as well as the control of the motion in the space. In our proposal, we suggest the use of three identical water-jet positioned in a circular arrangement. Since they are identical, the experimentation of only one is enough to understand the behavior and with the superposition of the results, it is possible to predict the behavior of the three water-jet together.
Specification of water-jet washing machine
AC 100V 50/60 Hz
Regular discharge pressure
Approximately 6.0 MPa
Regular discharge water
Approximately 270 l/h
Approximately 4 kg
The measurement of the pressure at the two ends of the hose is done via the VESV/VESI series VALCOM pressure transmitters. Meanwhile, the mass flow rate is calculated indirectly through the measurement of the of the mass of the water that is expelled from the outlet and collected inside a container. The mass of this container is measured with a AJ series Vibra balance.
Mass flow rate measurement
The pressure drop of the water varies depending on the state of the hose. Consequently, the pressure drop in the hose does not have a steady value. Since it is not feasible to obtain a unique friction factor for the hose, we aim to calculate the highest possible value. In order to calculate this value, most of the hose was gathered in a coil of approximately 0.15 m of diameter. We consider that this condition generates one of the highest friction factors that the system could present due to the continuous changes of direction of the flow of water. These changes generate secondary flows along the hose and these augment the friction. The only moment this condition is presented is when the device is collected before being deployed. During the deployment, it is unlikely that this situation happen. Therefore, it is unlikely to generate larger pressure drop values than those calculated on this condition.
Thrust force measurement
In order to generalize the results of the thrust force regardless the pump used, this force is described in function of the mass flow rate as is shown in the Fig. 14. In this general description, second order functions have the best matching to describe the behavior. With this chart, it is easier to visualize that the increasing rate of the thrust force is larger the smaller the diameter of the nozzle. However, the smaller the diameter of the orifice, the smaller the maximum thrust force that can be reached. This is due to the mechanical limitation that the pump has, which is reached easily with the smallest nozzles because the pressure increases rapidly.
Summary of water-jet experiments
After the experimentation with the test apparatus, we found out that despite that the reduction in the diameter of the orifice increases the speed of the fluid, the flow rate decreases at a larger rate and, consequently, the thrust force did not increase as expected. Furthermore, the main limitation for the thrust force is the maximum pressure that the pump is able to bear. The pressures developed are relatively large, however, they generate a reduced part of the thrust force. Meanwhile, the flow of water at the outlet is the main source for the thrust generation. It could be considered that a large flow with small pressure is preferable. In the analysis of the system, we found this characteristics with the nozzle of 0.8 mm and this diameter was chosen for the prototype. Furthermore, with the nozzles of diameter smaller than 0.5 mm, it was necessary to use filtered water since the tap water contains small particles that block the nozzle. Practically speaking, larger diameter is preferable to avoid this problem.
A larger diameter could further reduce the pressure, however, this may not bring enough advantages since the thrust force is not going to increase any further due to the limitation of the pump, as shown by its behavior in Fig. 13. Moreover, considering the behavior of the mass flow rate shown in Fig. 8, an incremental in the consumption of water is expected. Another possible way to optimize the thrust force is by increasing the flow without increasing the pressure is avoiding abrupt changes of direction in the nozzle and with a gradual reduction of the output instead of the orifice. Even so, these changes require a complex manufacturing in a reduced space, that could make it not feasible.
Finally, we found that despite the reduction in the diameter of the orifice, the speed of the water and consequently the water-jet thrust force did not increase as expected because the resistance factor incremental reduces the mass flow rate. Furthermore, this resistance incremental makes that the pressure elevates faster and the mechanical limitations of the pumps are reached faster as well.
With the three water-jets controlled independently, it is possible to generate spatial motion. However, as happens in all the underactuated systems, a sequence of movements are necessary to reach some points, since it is not possible to reach them with direct motions. In particular for our configuration, due to the disposition of the water-jets nozzles, there is not any control on the rotation on the longitudinal axis of the probe. This necessitates either the addition of a mechanism that allows fixing of the orientation of the nozzles relative to a fixed reference or a sensor that acquires the attitude of the nozzles in order to do the corresponding compensation depending on the current orientation.
Four main elements for the electronic design were considered: user interface, sensing, processing, and power stage. Taking into account that a straightforward control is desired, a three axes joystick (Sakae 30JH) is used as the interface for the user. The three signals acquired from the joystick allows defining the direction of the movement, as well as the magnitude of the thrust force.
Considering the lack of control in the longitudinal axis and that any kind of mechanism to fix the orientation of the nozzle to an inertial frame will increase the dimensions of the device, we decided to use an inertial-measurement unit (IMU) at the tip in order to acquire the attitude at any moment. In particular, the IMU used has a 3D accelerometer, 3D magnetometer as well as three-axis gyroscope. Even though the calculation of the attitude is possible by using the data from only the gyroscope, we decide to integrate the data of all the sensors embodied in the IMU to have more precise values. Furthermore, depending on the requirements of the task, other sensors can be added. For instance: thermometer, camera, dosimeter, and so on.
Because the location of the IMU is far from the base of the system, it is better to perform the processing of its information nearby the tip. Furthermore, taking into account that the data needs to be sent to the base as well as protected as much as possible from electromagnetic noise, one of the best options to transmit the data is through CAN bus. Therefore, a microcontroller is necessary at the tip to process the information from the IMU and for the adaptation of the resultant data to be sent by CAN protocol. For that reason, we selected the HiBot TiTech M4 Controller board, which possess all the required elements mentioned before. At the base of the probe, another microcontroller is required to receive the information from the TiTech M4 and process the attitude data together with the information of the desired direction of motion sent by the user through the joystick. For this task, we chose the HiBot TiTech SH2 Controller board.
In regard to the control of the whole device, we decide to adopt a control similar to the radio-control cars, since, like them, our system is underactuated and a forward motion is generated in order to move sideways. The main difference is that our device can move in three-dimensional space. The x-axis from the joystick represents the horizontal motion, meaning left-right motion, and the y-axis represent the vertical motion, meaning the up-down motion. This two motions are defined on a local reference frame that is located at the tip of the probe, which moves all the time with the probe and keeps its y-axis perpendicular to the ground of the user: the inertial reference frame. Finally, the z-axis of the joystick controls the magnitude of the thrust force. As a result, through the attitude obtained via the IMU, the flow of the pumps is adapted automatically to any orientation that the probe has, in order to fulfill the motion in the wanted direction. We also chose this strategy of control because up to this point in the design, there is no camera installed on the probe and the operator needs to keep sight of the position of the probe. Later, when a camera is installed at the tip of the probe, this control can be modified in order to define all the motions completely with respect to the local reference system located at the tip of the probe.
Tests were conducted in different environments to simulate more typical conditions such as: on the ground (pavement), in air, and in water.
The usefulness and efficiency of the IMU to correct the lack of control in the longitudinal axis was more notorious in the experiments carried out in the air and inside the water. The reason is that, in these environments, the device has less restriction to rotate and lose its reference. The IMU also helped to more easily control the system because the orientation of the reference frame at the tip was partially adapted to the reference of the user.
In general, the use of the IMU to acquire the orientation at the tip as well as the delimitation in the thrust force field allows us to control the system in different environments in a straightforward way. Nevertheless, the motion on the ground requires the implementation of snake-like motion, which is necessary to overcome the limitation generated by the large friction force. Another way to overcome this limitation is by using more powerful pumps since the pumps used in the experiments belong to middle-range-pressure washing machines aimed at home use. On the other hand, in the tests done with the probe in the air, we were able to keep the tip gliding in the air. However, the control was extremely sensitive and just a small change of direction in the joystick caused an impulsive motion at the tip. This made it very difficult to direct the probe to a fixed point due to the lack of drag force. A possible way to solve this problem is by the implementation of a PID control in order to damp the impulsive motion that is generated when the probe is in the air. The experiment in the pool helps us to understand that in order to make it easier to control our system, an extra degree of freedom that pulls the probe backward may be useful. This degree of freedom could be added by the addition of a motorized reel which collects the probe when it is necessary. Finally, although the developed probe is larger than the grating hole due to the size of the IMU, we can minimize its size by custom made in the near future. In the current device, we use a commercially available IMU for ease implementation.
In this paper, we propose the usage of water-jet as propulsion source with the aim of solving the need of developing more slender and longer devices for inspection tasks. An analysis of the variables that interact in the generation of thrust force was developed with the intention of understanding their correlation. We found that the main limitation for the thrust force is the maximum pressure that the pump is able to bear. A first prototype was developed to test the mobility and controllability of the device on the ground, in the air and under water. It was possible to move and control the device on the ground, in the air and under the water. However, in the aquatic environment, the system shown the best controllability and mobility. Finally, the water jet seems promising to be used as propulsion source for inspection systems, however, further research is necessary to improve and explore different possibilities of this concept.
All authors equally contributed. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Moore T (1985) Robots for nuclear power plants. Int Atomic Energy Agency Bull 27(3):31–38Google Scholar
- Seward DW, Bakari MJ (2005) The use of robotics and automation in nuclear decommissioning. In: 22nd international symposium on automation and robotics in construction, vol. 44. ISARC, Ferrara, ItalyGoogle Scholar
- Roman HT (1991) Robots cut risks and costs in nuclear power plants. IEEE Comput Appl Power 4(3):11–15. doi:10.1109/67.85957 View ArticleGoogle Scholar
- Fischetti MA (1985) Robots do the dirty work: Some walk, some roll as they go about their tedious, hazardous chores in nuclear-power plants. IEEE Spectr 22(4):65–73. doi:10.1109/MSPEC.1985.6370621 View ArticleGoogle Scholar
- Sugisaka M (2011) Working robots for nuclear power plant desasters. In: 5th IEEE international conference on digital ecosystems and technologies (IEEE DEST 2011), vol 5, pp 358–361. IEEE, Daejeon, South Korea. doi:10.1109/DEST.2011.5936593
- TEPCO: TEPCO application of robot technology. http://www.tepco.co.jp/en/decommision/principles/robot/index-e.html. Accessed 10 Sept 2016
- Tokyo Electric Power Company: Investigation results of inside the Fukushima Daiichi Power Nuclear Plant Unit 2 Primary Containment Vessel (PCV)-Guide Pipe removal-. Technical report, TEPCO (2013). http://www.tepco.co.jp/nu/fukushima-np/handouts/2013/images/handouts_130422_06-j.pdf (in Japanese)
- Robinson G, Davies JBC (1999) Continuum robots—a state of the art. In: Proceedings 1999 IEEE international conference on robotics and automation (Cat. No.99CH36288C), vol 4, pp 2849–2854. IEEE, Detroit, Michigan, USA (1999). doi:10.1109/ROBOT.1999.774029
- Rolf M, Neumann K, Queißer JF, Reinhart RF, Nordmann A, Steil JJ (2015) A multi-level control architecture for the bionic handling assistant. Adv Robot 29(13):847–859. doi:10.1080/01691864.2015.1037793 View ArticleGoogle Scholar
- McMahan W, Chitrakaran V, Csencsits M, Dawson D, Walker ID, Jones BA, Pritts M, Dienno D, Grissom M, Rahn CD (2006) Field trials and testing of the OctArm continuum manipulator. In: Proceedings 2006 IEEE international conference on robotics and automation, 2006. ICRA 2006, pp 2336–2341. IEEE, Orlando, Florida. doi:10.1109/ROBOT.2006.1642051
- Buckingham RO, Graham AC (2010) Dexterous manipulators for nuclear inspection and maintenance—case study. In: 2010 1st international conference on applied robotics for the power industry (CARPI 2010), pp 1–6. IEEE, Montréal, Canada. doi:10.1109/CARPI.2010.5624476
- Mehling JS, Diftler MA, Chu M, Valvo M (2006) A minimally invasive tendril robot for in-space inspection. In: The first IEEE/RAS-EMBS international conference on biomedical robotics and biomechatronics, 2006. BioRob 2006, vol 2006, pp 690–695. IEEE, Pisa. doi: 10.1109/BIOROB.2006.1639170
- Walker ID (2013) Robot strings: long, thin continuum robots. In: 2013 IEEE aerospace conference. IEEE, Big Sky, Montana, pp 1–12. doi:10.1109/AERO.2013.6496902
- Buckingham R, Graham A (2012) Nuclear snake arm robots. Ind Robot Int J 39(1):6–11. doi:10.1108/01439911211192448 View ArticleGoogle Scholar
- Horigome A, Yamada H, Endo G, Sen S, Hirose S, Fukushima EF (2014) Development of a coupled tendon-driven 3D multi-joint manipulator. In: 2014 IEEE international conference on robotics and automation (ICRA). IEEE, Hong Kong, China, pp 5915–5920. doi: 10.1109/ICRA.2014.6907730
- Fukuda J, Konyo M, Takeuchi E, Tadokoro S (2014) Remote vertical exploration by active scope camera into collapsed buildings. In: 2014 IEEE/RSJ international conference on intelligent robots and systems. IEEE, Chicago, Illinois, USA, pp 1882–1888. doi:10.1109/IROS.2014.6942810
- Shinsho: Shinsho Nozzles. http://www.ss-shinsho.co.jp/nozzle.html. Accessed 16 Sept 2016
- Fricke: Lindauer Schere. http://www.fricke-kanal.de/Leistungen/TV-Inspektion/Lindauer-Schere/lindauer-schere.html. Accessed 16 Sept 2016
- Jetlev-Flyer. http://jetlev-flyer.com/. Accessed 16 Sept 2016
- Zapata Racing. http://zapata-racing.com/en/. Accessed 16 Sept 2016
- Wu G, Sun H, Zou J, Wan L (2009) The basic motion control strategy for the water-jet-propelled USV. In: 2009 international conference on mechatronics and automation. IEEE, Changchun, China, pp 611–616. doi:10.1109/ICMA.2009.5245049
- Guo S, Lin X, Hata S (2009) A conceptual design of vectored water-jet propulsion system. In: 2009 International conference on mechatronics and automation. IEEE, Changchun, China, pp 1190–1195. doi:10.1109/ICMA.2009.5246462
- Mazumdar A, Lozano M, Fittery A, Asada HH (2012) A compact, maneuverable, underwater robot for direct inspection of nuclear power piping systems. In: 2012 IEEE international conference on robotics and automation. IEEE, Saint Paul, Minnesota, USA, pp 2818–2823. doi:10.1109/ICRA.2012.6224619
- Tokyo Electric Power Company: Investigation result of inside of the Primary Containment Vessel ( PCV ), 1F-2. Technical report, TEPCO (2012). http://www.tepco.co.jp/en/nu/fukushima-np/roadmap/images/m120123_01-e.pdf
- Tokyo Electric Power Company: Summary of Decommissioning and Contaminated Water Management. Technical report, TEPCO (2015). http://www.tepco.co.jp/en/nu/fukushima-np/roadmap/images/d151001_01-e.pdf
- Tokyo Electric Power Company: Development of a technology to investigate inside the Reactor Primary Containment Vessel ( PCV )-Site test “Investigation B1” on grating around the pedestal inside Unit 1 PCV-. Technical report, TEPCO (2015). http://www.tepco.co.jp/en/nu/fukushima-np/handouts/2015/images/handouts_150406_01-e.pdf