For a long time, this type of propulsion was normally used because of its comfortable design and its great propulsion with higher weights. In , this type of propulsion was used in the development of a torpedo-shaped robot called TSL, where the bow and stem propulsion systems had a tunnel for the performance of water jets [ 48 ].
In spherical-shaped AUVs, this type of drive is used in a vertical direction; two actuators can be controlled by one thruster and one servomotor; the jet-based thruster decides the value of the driving force, and the servomotor controls the height of the thruster [ 37 , 51 ].
Nessie III and DaryaBird used the same brush motor-based propeller offered by the SeaBotix brand, with a consumption of W and an ability to withstand depth up to m; these were used in directional movement in the -xy plane; in the case of the second robot mentioned, it used a RoboPlus Hibikino thruster with a power of 90 W for the robot descent control.
Brushless motors became popular in small and medium ROVs; their use dates back to the 80s and 90s. An example is ABE, an AUV destined for benthic species exploration that uses brushless motors brushless with oil [ 52 ] for pressure compensation.
Another related example is the operation of Jeff, a small AUV designed for inspection and swarm joint work; the propulsion system has mainly two DC motors with a custom magnetic coupling design to avoid corrosion and short circuits. In the design and development of UUVs, a new structure was chosen for the steering management in the 6-DOF; this particular configuration in parallel can be seen in Figure 38 ; it has two main thrusters in the front and another in the rear that handles the steering of the robot [ 54 ], and the union between the two parts is through the hydraulic system.
The functions performed by the UUVs include the inspection, manipulation, and data collection; all the robots have a video transmission system implemented, so it is necessary to develop a lighting system to acquire images since the underwater environment does not have visibility conditions due to the lack of a light source. In another application in image acquisition, there are the recognition and detection of objects, for rescue or supervision robots, and a clear example of detection is found in the research of Pedersen et al.
This meaning applied to robotics results in an automatic handling machine, reprogrammable in either a moving or fixed position. In , a hybrid underwater robot was made, with a crab and lobster structure, where its legs acted as manipulators and its main function was to inspect underwater structures and shipwrecks in shallow waters, where activities such as cable cutting, grinding, and drilling are required [ 56 ]. Table 3 summarizes all the aquatic robots seen in this section.
Navigation, in simple terms, conforms to particular methods that allow someone to know where they are and how to get to a new point. Depending on the type of environment and available reference points, these methods could be simple; however, they could become complex results in a hostile, changing, and unpredictable environment, also, to reference points that are not visible [ 6 ].
The tasks that the UUVs must perform require navigation to displace to different location points to complete their duty. Usually, the tasks performed by ROVs demand heavy work and can be confined to smaller spaces where navigation is probably not very complex and could be performed directly by the operators through a joystick and the use of one or more video cameras installed on the ROV.
Moreover, autonomous vehicles have more tasks and must carry out missions that take several hours, days, or even months [ 57 ]. Most of these missions are focused on conducting maritime exploration which covers large areas of several hundred square kilometres.
Navigation plays a vital role here that if it is not properly executed, it could not only affect the fulfilment of the mission but also affect the safety of the vehicle [ 11 ] and, in the worst scenario, could lead to the loss of the robot, causing economic loses and contamination of the explored environment.
Due to its autonomy, navigation must be accomplished under the control of a computer embedded in the vehicle. Navigation in AUVs represents a great challenge for most researchers due to the impossibility of using a global positioning system GPS underwater.
The electromagnetic radiation waves emitted by satellites are absorbed when they come into contact with water, so a GPS signal receiver cannot capture the waves underwater [ 58 ]. Therefore, underwater devices have been used to establish a local positioning system.
Thanks to advanced technology, many of these devices have been improved and optimized considerably in terms of dimensions and performance. This motivated the investigation and improvement of methods that allow a better estimation of the location and thus more exact navigation.
If the travel speed of the AUV is known, new positions can be estimated by consecutive integrations of speed. To perform velocity measures, Doppler velocity log DVL is normally used in conjunction with inertial systems and a compass; the estimated position solely depends on the movement of the vehicle and hence the proprioceptive name. This type of methodology is known as dead reckoning [ 59 ]. This type of navigation system corresponds to the research written by Itzik Klein and Roee Diamant; they developed a system that estimates the trajectory travelled by a water vehicle that moves freely in the direction of the marine currents [ 60 ].
Because these robots work very closely to the sea surface, they are easily susceptible to orientation change, which creates problems in the path. Another interesting work about accelerometers and how they are used corresponds to the authors Yan et al. They worked on a dead reckoning navigation system based on neural networks using only accelerometers, due to the cost of using other sensors such as a DVL or the dependence on an acoustic system.
The errors that have been generated by using inertial units are reflected in rapid changes in the measured angles by gyroscopes; this considerably increases the error of the dead reckoning system. Normally, the proprietary navigation systems work in conjunction with external systems to correct themselves and reduce the accumulated error.
Some of the work done on dead reckoning which works in conjunction with other reference systems corresponds [ 61 ] to Kepper et al.
Due to the noise generated by the IMU, the raw data captured was filtered using an extended Kalman filter. For the model effectiveness, implementation was evaluated in data collected in 3 different environments for field experiments and in an open ocean environment. Correct navigation requires a good position estimate, so the instruments and the algorithms used must obtain the most exact ubication.
The main problem of proprioceptive navigation is that the error increases limitlessly as the distance travelled by the vehicle increases. If an external reference system is not considered, the navigation becomes critical for the vehicle and its mission.
As a solution to this problem, acoustic navigation is employed. Acoustics waves are appropriate for underwater propagation due to minimal attenuation. Hence, they are employed for underwater communication and positioning the underwater vehicles.
For example, underwater vehicles use data of the placed beacons for estimation of their positions in the work zone.
The standard LBL method is characterized by beacons or transborder, which is fixed as shown in Figure The image shows the configuration system for the vehicle and transponders. First, transponders listen to the pings emitted for the vehicles, and distance estimation is obtained from TAT turnaround time at a specific frequency. Thus, the vehicle can estimate its position by algorithms based on recursive least squares RLS or using extended Kalman filters. The vehicle must save transponders positions.
Those who developed REMUS [ 8 ], a torpedo-shaped underwater robot created for exploration of marine resources, presented two modes of operation: autonomous and nonautonomous.
In the autonomous mode, the robot had to follow a path formed by acoustic transponders implanted on the seabed, and REMUS acted as a target hunter.
The transponders distribution defined the navigation path of REMUS; on the other hand, in the nonautonomous mode, the navigation was carried out with the help of a boat, and REMUS followed it through an acoustic communication. This method works quite well for a single vehicle. In the case of the navigation with several vehicles, variants of this method have been proposed to eliminate the consultation signals, converting the communication in one direction and thus removing the dependence of the time intervals for vehicle location updates.
However, both the beacons and the underwater vehicles must be synchronized [ 63 ]. The standard configuration of the LBL system and its variants Figure 41 allows establishing an absolute positioning for either one or several underwater vehicles.
However, the task of implementing and calibrating polygonal beacon arrangements is expensive and difficult. Therefore, it was decided to improve these systems even more and only one beacon has been achieved to determine the position of a vehicle; this system has been called Ultrashort Baseline USBL.
This configuration is illustrated in Figure This type of configuration works similarly to the standard LBL configuration; however, the vehicles have multiple acoustic receivers, because they must determine not only the distance at which they are from the beacon but also the angle with which the replica of the signal arrives. The query signal was issued. In this way, the need for using several beacons for the trilateration calculation is avoided.
We can cite the work done by Hidaka et al. They implemented this acoustic navigation system which intended to use an array of hydrophones that were very close.
The angle was calculated to the arriving sound from the offset that occurred between the hydrophones. Optical navigation uses optical devices such as video cameras or optical diodes from which morphological data of the seabed are recorded. Carreras et al. In the work, the location algorithm details through some graphic results and the precision of the system. The algorithm allows obtaining a 3D position, orientation, and speed of the vehicle by detecting reference points from the bottom of the tank.
The location estimates are highly accurate without drift, allowing them to be used as feedback measurements for low-level speed-based controllers. Its computing system is It is necessary to take controlling the orientation and movement of underwater vehicles into account, and it may demand an exploration mission or some work that requires manipulation or extraction on the seabed.
However, due to the presence of external disturbances and uncertainties in the marine environment, linear control methods are not very efficient, so it is necessary to apply advanced robust control methods.
The objective of an orientation control is to retain the required orientation regardless of swell and unpredictable disturbances in the environment.
That is why a hydrodynamic model and mathematical parameters of the structure must be obtained first. Likewise, to establish a control system, the following points must be taken into account: the performance of the system is limited, adding that the behavior of the control system must be robust in terms of both stability and performance, since it takes into account the energy management and optimization of the entire system [ 65 ].
This approach takes into account the inevitable imperfection in physical systems and variables; one of the investigations on the performance of a new control strategy for imperfect systems is observed in [ 66 ], starting from an electromechanical system based on a light structure that acts as a support and supply for the simple coils found in the structure; the purpose of this research is the simulation of control systems for imperfect systems that, thanks to the peculiar properties in the structure, the effects of vibration signals on the hidden dynamic system of the imperfect system can be observed.
Given the premise on nonlinear control systems in imperfect systems with more than two variables, it is considered that most of the research carried out within the field of hydrodynamics and the behavior of an ROV is established in only the movement controls, guaranteeing the movement of the robot in the established route without considering the dynamics hidden in the structure.
However, most researchers make use of Computational Fluid Dynamics CFD simulations on the behavior of their framework to reduce the error due to changing environmental conditions which are difficult to predict.
The sliding mode control SMC is a robust control for modelling uncertainty and parameter variations and has good disturbance rejection characteristics. There have been a wide variety of applications of the same [ 67 — 71 ]. However, it inherits a discontinuous control action; therefore, the chattering phenomenon that occurs when the system operates close to the sliding surface will occur. Sometimes this discontinuous control action can even make system performance unstable.
Side Zhao and Junku Yuh proposed an adaptive control based on a disturbance observer [ 69 , 72 ]; the control scheme of this system has an adaptive controller based on a nonregressor, and it is the outer loop of the control scheme, while the inner loop controller is the disturbance observer.
These two elements mentioned above are the components of the adaptive control system proposed by the authors, which is robust against external disturbances and unpredictable behaviors due to the self-adjustment of its control parameters. Recently, neural networks have gained considerable attention in robotic systems control due to their versatile properties, such as nonlinear mapping, learning ability, and parallel processing [ 67 , 69 , 73 ].
The most useful feature of neural networks in control is their ability to approximate arbitrary linear or nonlinear mapping through learning. Due to this property, neural networks have been proven to be a suitable tool to control complex nonlinear dynamic systems. However, due to their arithmetic complexity, their implementation in engineering is not easy. Control based on fuzzy logic or fuzzy control FC in English is a control that has supplanted conventional technologies in many applications [ 35 , 68 , 74 — 76 ].
An important property of fuzzy logic is its ability to express ambiguity in human thought. Therefore, when the mathematical model of the process does not exist or does exist with uncertainties, the FC becomes an alternative way for dealing with the unknown process. However, the large number of fuzzy rules for high-order systems makes the analysis complex. A fuzzy-based depth control scheme is illustrated in Figure 44 and a fuzzy-based yaw angle control scheme is illustrated in Figure Table 4 summarizes the main contributions of some additional navigation and orientation control methods that correspond to those most used by UUVs.
In the first row of Table 4 , some linear methods of proportional-integral-derivative PID type are included, which work in conjunction with the other previously reviewed methods.
The ROVs first shape was rectangular with an open hull and positive buoyancy; it was so big that it could not be transported by a single person, and it was necessary to place a pulley in the water. Inside the investigated robots, it was found that only four thrusters provide 5 degrees of freedom compared to others that need six thrusters to reach 5 degrees. The studied robots determine that, to achieve all the degrees of freedom, the robots must have eight thrusters installed or five thrusters with two actuators to change the force direction.
One of the most used materials in the manufacture of aquatic robots is aluminum, because it does not deform at high pressures; it is a dense material and is not corrosive. Most of the researches evaluated are designed to operate under positive buoyancy; to be able to submerge, they must activate the immersion thrusters; and to return to the surface it is enough that the thrusters are deactivated. The improvement of aerodynamic and hydrodynamic characteristics of aquatic robots with a biomimetic approach has accumulated great results, improving very important factors in the design such as hydrodynamic drag, propulsion force, and energy consumption, giving room to achieving better results with further study.
It is important to mention that the biomimetic form of a robot not only implies improvements for itself but also reduces the degree of risk to possible alterations to a natural biological environment at the time of the interaction. The constant improvement of biomimetic technology has broken the trend of only implementing robots based on propulsion by caudal or pectoral fin; studies have opened a new window for the use of intelligent actuators, materials capable of providing better mechanical characteristics, such as greater flexibility under specific conditions, getting closer to the efficiency of real biological models with diverse morphological characteristics.
The use of a pressure sensor has become much more standardized in the manufacture of any UUV, simply to obtain the depth data. However, some of these robots still have a dedicated depth sensor, thus achieving a greater comparison range between points. In most current ROVs, we can observe the constant use of an IMU sensor with the combination of sonar to find the underwater positioning, also applying a filter for the correct interpretation of data.
GPS modules are used more in AUVs than in ROVs, because ROVs present a physical connection between the robot and controller, while the AUVs are programmed with a path or route to follow; that is why they emerge to the surface to obtain their position before making a submersion. The use of brushless motors has become very popular with the integration of propellers. It is found in different types of UUVs long before the 20th century.
The advantage of this type of motor is adequate cost, better quality, and less maintenance than other motors. The application of new sensors for the acquisition of oceanographic data in robots has become increasingly common, as a result, mainly due to the growing interest in the study of marine ecosystems and the conservation of species.
Many of the works reviewed, related to the control of direction or displacement with different engines, do not show much detail in the electronic components used, making it difficult to trace an evolutionary timeline of emerging technologies of electronic components used in UUVs.
You can see the trend towards map-based navigation methods as opposed to those that use fixed beacons around their exploration environment. The orientation and movement control is applicable for both ROVs and AUVs, highlighting the routes control and trajectory tracking towards autonomous vehicles. The trend of new control methods is to apply combinations of more than one method to improve their characteristics and achieve finer control. This is an open access article distributed under the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Article of the Year Award: Outstanding research contributions of , as selected by our Chief Editors. Read the winning articles. Journal overview. Special Issues. Academic Editor: L. Received 21 Feb Accepted 07 Jun Published 06 Jul Abstract Since its beginning, around the 50s decade, until present days, the area of unmanned underwater vehicles UUV has considerably grown through time; those have been used for many tasks and applications, from bomb searching and recovery to sea exploration.
Introduction Robotics is a branch of engineering that involves the concept, design, manufacture, and operation of programmable machines, which can develop their autonomy in the execution of a specific task. Figure 1. Location of unmanned water vehicles within a basic robotics classification. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure Mechanical design of a rectangular ROV [ 20 ]. Underwater robot for shallow depth [ 21 ]. Classification of bioinspired robots by their swimming mode [ 22 ].
Underwater manipulator design [ 28 ]. Bioinspired alternative shell and drive design [ 29 ]. Table 1. Table 2. Total ATP analyzer with a microfluidic device [ 42 ]. Construction of external orthogonal coordinate system [ 49 ]. Direction change operation diagram [ 54 ]. Table 3. Variants of the standard LBL configuration.
Diagram of the adaptive control system based on a disturbance observer [ 72 ]. Fuzzy control system diagram for depth [ 74 ]. Fuzzy control system diagram for guidance [ 74 ]. Fuzzy i Avoid collision in marine vessels through an intelligent decision-making system [ 83 ].
Adaptive i Stabilize the motion control of an AUV disturbed by unknown hydrodynamic coefficients [ 87 ]. Sliding modes i Improved response, insensitive to parameter variation and disturbance [ 90 — 93 ]. Neural networks i A bioinspired neurodynamic model is presented, used for a kinematic controller [ 94 ].
Table 4. Classification of orientation and movement control methods most used in UUV. References J. Cohen, C. Small, A. Mellinger, J. Gallup, and J. View at: Google Scholar E. View at: Google Scholar S. Chutia, N. Kakoty, and D. View at: Google Scholar B. Sahu and B. View at: Google Scholar R.
Wynn, V. Huvenne, T. Le Bas et al. View at: Google Scholar C. Allen, T. Austin, and R. Allen, R. Stokey, T. Austin et al.
View at: Google Scholar M. Radojevic, M. Nawaf, F. Maurelli et al. Xiang, C. Yu, and Q. View at: Google Scholar X. Ai, S. Kang, and W. View at: Google Scholar G. Yao, J. Liang, T. Wang et al.
Lin and S. Note that we implicitly is a function of the vertical position. In particular, substituting 4. Moreover, if the tip of the actuator is constrained, the following relationship between input pressure, bending angle, and bending force can be obtained from Eq. Interestingly, since Eq. Experimental setup. The actuators were fabricated using a multi-step molding process with 3D printed molds [22, 40].
Woven fiberglass S plain weave was attached to the flat bottom surface using silicone adhesive as the strain limiting layer. Kevlar fiber of 0. Fiber reinforcements were further secured by placing the entire assembly into another mold to encapsulate the actuator body in a 1. The actuator body was then removed from the steel rod Fig. A vented screw was fed through the silicone cap and combined using the dimensions of the extension bar to obtain became the connection for the pneumatic tubes.
A sensor layer the overall force at the actuator tip regardless of its direction. The sensor layer had a 6xmm model were determined by calibration. Three actuators were central pocket for a flexible bend sensor Spectra Symbol Flex used in this study, with the same actuator design and the same Sensor FS-LST of 5xx1mm. The sensor material for their top and bottom walls, therefore we used exhibited very good linearity 0.
The same calibration procedure was followed to obtain resistance when calibrated with a goniometer. Despite using a different actuator design of , , , An experimental setup was developed Fig.
The reported in [10, 22]. The proximal tip of the actuator was same calibration procedure was used again to estimate the mounted firmly to the mounting base and connected to the air sensor stiffness 0. Free-space bending test direction with respect to the horizontal surface. To adapt to the In the free-space bending test, the pressure-angle relationship variation of contact direction, an extension bar was attached to of 4.
In the experimental c IEEE. Isotonic test results. Comparison of force predictions from the analytical force model Analytical and experimental force measurements Exp.
Isometric test results. Isometric test In an isometric test, the actuator was constrained at constant setup, an actuator was mounted horizontally on the platform, bending angles while input pressure increased from 0 to with the distal tip bending downwards in the vertical plane kPa. In each state the contact forces were measured by the without obstacles. Without rigid structural support, the mounted force sensor. The experiment consisted of three trials, with three actuator exhibited a degree downwards natural bending actuator positions considered for each trial: 90, , and angle due to gravity.
The analytical model was used to was pressurized from its natural resting position with air estimate bending forces using the measured bending angles. To validate the analytical model of 4. To illustrate the 8 d.
The analytical results form parallel lines for each bending influence of the integrated bend sensor, two pressure angle. This trend is matched by the FEM and experimental estimations were made by setting 0 and 0.
They are compared in Fig. The analytical model was derived and hence FEM element deformations. The experimental based on internal material stretch, which should be zero at the measurements are closely matched by the analytical results with natural resting condition. In the results shown in Fig. Isotonic test match those obtained experimentally throughout the actuator In the isotonic test, the input pressure was kept at a range of bending range.
Moreover, considering the sensor stiffness constant values, and interaction forces were measured at results in a noticeably better match to the experimental different bending angles. The The experimental results are shown in Fig. The justified and used in the rest of this study.
The intersections with the vertical axis indicate friction of fiber braiding and between different layers of the maximum achievable force for that pressure, which occurs materials within the actuator [6]. However, although our at zero bending. Between the two intersections, the force-angle actuator design utilizes a fiber layer, there is minimum sliding relationship is nonlinear as described in 4.
This between the fiber and the actuator body, and hence the friction nonlinearity is introduced by the hyperelastic material property loss is negligible during actuation.
Therefore the dominating and therefore captured very well by the analytical model. The internal interaction is material stretch, which does not lead to comparison results for each pressure configuration are listed in hysteresis, as observed in previous work on soft bending TABLE I. The analytical model provides a better estimation for actuators made from the same soft material [34].
In our previous larger pressures, with 2. This is mainly due to the c IEEE. The results demonstrated that the analytical model 60 1. In addition, the 3. Potential applications include complex actuation procedure the actuator underwent not fully glove-type wearable robotic devices where interaction forces considered in the analytical model, such as interactions between are mostly encountered at the actuator tip [10, 23], and new different material layers, nonlinear bulging and deformations actuator designs featuring the same topology.
The above factors were In future work, analytical modeling of nonlinear actuator more significant for lower pressures, where the actuator started behaviors, such as non-uniform bending and radial bulging will deforming from its original state. Although an averaged be tackled. Viscoelastic effect, air compressibility and heat material shear modulus was used to incorporate the above dissipation arising from fast actuation will be investigated to factors, the calibration process for the shear modulus resulted analyze the dynamic actuator behaviors.
Distribution of in an optimal value for the entire pressure range from 0 to interaction forces along the actuator body will be investigated, kPa. Therefore, the resulting analytical model provided together with engineering aspects such as actuator fatigue and better predictions for higher pressures.
The isotonic test results failure modes. The modeling approach could be generalizable also illustrated the compliant feature of soft bending actuators, to other actuator designs following the same fundamental where each isotonic line in Fig. Applications of soft bending actuators will be carried out exploring the integrated bending sensor, in scenarios such as VI. Through modeling analysis and experimental validations, the static and dynamic properties This work demonstrated that the bending and tip-force of the actuators could be demonstrated to the robotics research capabilities of the soft fiber-reinforced bending actuators under community.
With its unique features and the tools for analysis, quasi-static conditions could be characterized, modeled, and design, and control, soft bending actuator will be a capable controlled. Moreover it was shown that integrating an off-the- competitor of rigid-bodied actuators for robotic applications.
Although the study assumed [1] Haddadin, S. The quasi-static states with slow motions, its conclusions are International Journal of Robotics Research, An analytical model was developed to describe the force [4] Majidi, C.
Soft actuator the input pressure, bending angle and tip bending force. The mimicking human esophageal peristalsis for a swallowing resulting force model consists only of polynomial functions, robot. Pneumatic artificial muscles: actuators for robotics and automation, European journal of Mechanical real-time calculation and control in robotic applications. The and Environmental Engineering, 47 1 : 11—21, A [7] Wait, K.
A pneumatically actuated geometrical study was conducted for the actuator on the quadrupedal walking robot. A nonlinear controller for pneumatic To validate the analytical models, a FEM model of the actuator servo systems: Design and experimental tests.
Mechatronics, was developed closely resembling the real actuator. High-accuracy tracking match to the reality, while compromising real-time capability. Distributed Computation. Soft robotic glove for combined assistance and at-home Sukhatme Ed. Springer Tracts in Advanced Robotics, DOI: A novel monolithic piezoelectric Signal has very high frequency e.
There are certain design criteria that are desired when building a servo motor, which enable the motor to more adequately handle the demands placed on a closed loop system. First of all, servo systems need to rapidly respond to changes in speed and position, which require high acceleration and deceleration rates.
This calls for extremely high intermittent torque. Servos As you may know, torque is related to current in the brushed servo motor. So the designers need to keep in mind the ability of the motor to handle short bursts of very high current, which can be many times greater than the continuous current requirements. Another key characteristic of the brushed servo motor is a high torque to inertia ratio.
This ratio is an important factor in determining motor responsiveness. Further, servo motors need to respond to small changes in the control signal. So the design requires reaction to small voltage variations. Digital hydraulics, new! Practically In this class we will use only servos In past we used DC motors with H-bridge, pneumatic actuators, nintinol wires and hydraulic actuators. So far, if you want to build rather small robots and you want to concentrate on intelligence and sensing, RC servos are the best choice.
Many new types arrive every year, from very small to big powerful ones. Look to internet. We will learn about some new actuators if time will allow at the end of the class. Open navigation menu. Close suggestions Search Search. User Settings. Skip carousel. Carousel Previous. Carousel Next. What is Scribd?
Explore Ebooks. Bestsellers Editors' Picks All Ebooks. Explore Audiobooks. Bestsellers Editors' Picks All audiobooks. Explore Magazines. Editors' Picks All magazines. Explore Podcasts All podcasts.
Difficulty Beginner Intermediate Advanced. Explore Documents. Actuators in Robotics. Uploaded by Mike Lassa. Document Information click to expand document information Description: Manufacturing. Did you find this document useful?
0コメント