R&D PULSE
A GRIP IN THE DEEP
How equipping an underwater vehicle using a continuum manipulator with multiple degrees of freedom enhances its interventional capabilities.
Written by Mark Crawford
The light-duty and low-cost UVMS, equipped with a continuum manipulator. Image: Long Wang
FROM INSPECTIONS TO AQUACULTURE and exploration, underwater vehicle-manipulator systems (UVMS) have a number of practical applications. But researchers, scientists, and other users more often than not are looking for greater interventional capability. The way to do that is by integrating one or more manipulators onto free-floating underwater vehicles.
However, designing and controlling a free-floating UVMS is a challenge due to the physical design requirements, harsh environmental conditions, and unpredictable subsurface disturbances.
With the goal of taking the design and functionality of a UVSM to the next level, a research team at the Stevens Institute of Technology’s Department of Mechanical Engineering in Hoboken, New Jersey, have created a light-duty and low-cost UVMS that supports the use of a continuum manipulator for improved intervention.
Experimental approach
Robotics researcher Justin Sitler and professor Long Wang wanted to develop a kinematic model for a continuum UVMS so they could use that data to construct an algorithm to resolve the robot’s redundancy and generate joint space commands.
“Our manipulator design is a continuum manipulator, an unconventional class of robots characterized by high flexibility and compliance, allowing it to achieve continuous articulating profiles,” Sitler said.
This class of manipulators boasts several features that make them useful for underwater use. Natural passive compliance helps compensate for positioning error and minimize contact forces. And since continuum manipulators are fairly lightweight, most of their mass is in an actuation unit.
“This means inertial coupling between the manipulator and the vehicle is reduced, simplifying vehicle control,” Wang said.
Lastly, the continuum manipulator’s flexibility makes it particularly useful in navigating constrained environments, such as a natural crevice or a pipe.
The robot’s various components. Image: Long Wang
Research and design
This continuum UVMS features an open-source waterproof continuum manipulator and a free-floating underwater vehicle platform called BlueROV2. Provided by Blue Robotics, the platform was chosen for its low cost and ease of integration via an additional payload skid. It features customizable waterproof enclosures and penetrators, and a heavy configuration upgrade, which has eight thrusters to allow for six degrees of freedom control and improved stability.
Sitler and Wang experimented with different methods to optimize the continuum UVMS’s trajectory for specific tasks, using both a weighted least norm solution and the gradient projection method.
So, the researchers developed kinematics for the continuum UVMS and an optimized redundancy resolution algorithm for controlling its location via null-space gradient projection. Kinematic simulation data provided insights into the proposed algorithm’s performance via manipulator/vehicle control and optimization impact.
A two-segment multi-backbone design gives the continuum robot arm freedom of motion in four directions. This allows the distal portion of the arm to bend independently from the proximal portion, increasing its maneuverability and dexterity.
“The continuum manipulator is actuated by two sets of three nitinol wires, which pull on the end link of each segment,” Sitler said. “The wires are connected to an actuation unit, which creates the linear pulling motion via six lead screws and servo motors inside a waterproof enclosure.”
The continuum robot arm has a two-segment multi-backbone design, giving the entire arm four degrees of freedom. This allows the distal portion of the arm to bend independently from the proximal portion, increasing its maneuverability and dexterity. Image: Long Wang
To accomplish a number of tasks, the robot’s end-effector must achieve a specific pose (position and orientation), which requires the system to have six degrees of freedom. “However, a UVMS will typically have kinematic redundancy—meaning that with careful planning and decision-making, the robot can complete additional subtasks or objectives while still satisfying the primary task constraint,” Sitler explained.
Two sets of kinematic simulations validated the algorithm to resolve the free-floating vehicle system’s redundancy. The first set analyzed the weighting matrix’s impact and the second set analyzed the gradient projection’s impact. During both simulations, the robot moved the end-effector to a desired position and orientation, imitating a basic grasp of an object at a known location.
Control challenges
The most significant problem with the control system hardware was the lack of an effective navigation system. State-of-the-art systems that can provide real-time information are expensive and were inaccessible for this project.
“The onboard inertial measurement unit [IMU] could in theory be used for dead reckoning, but it is quite noisy and prone to drift,” Wang said. “Lacking adequate sensors or control for state feedback, we decided instead to implement UVMS control by only using open-loop control. The thruster gains were manually tuned to achieve the most accurate performance, but open-loop control fundamentally limits the accuracy of the experiments. Any future work will require development of a functional navigation system.”
“Better tuning of algorithm parameters may allow the robot to better balance speed and performance, as well as improve optimization of secondary objectives.”
—Long Wang, Assistant Professor, Department of Mechanical Engineering, Stevens Institute of Technology
The second shortcoming was the lack of a dynamic model for the UVMS, which made it difficult to control the potential sources of error.
“For example, the buoyancy of the UVMS in theory should be neutral,” Sitler said. “However, in practice, it is common to adjust the ballast so the vehicle is slightly positively buoyant in order to safely recover the vehicle in the event of loss of control. When the vehicle is given an upward velocity command, the buoyancy tends to add extra lift which causes greater acceleration than desired. Also, the inertial drift can cause the vehicle to not be fully at rest when initially executing the trajectory, which is a significant problem since we are using open-loop control.”
Future work should include improvements to the vehicle’s closed-loop controller to increase velocity accuracy and reduce end-effector positioning errors, which will be crucial for real-world applications.
“Better tuning of algorithm parameters may allow the robot to better balance speed and performance, as well as improve optimization of secondary objectives,” Wang said.
Developing new and more sophisticated tasks will be another important research area as the team aims to expand the system’s capabilities.
“Multi-manipulator planning may also be a lucrative research topic, with potential applications including dexterous underwater manipulation, perching, or climbing,” Sitler added.
Mark Crawford is a technology writer based in Corrales, N.M.
© 2025 The American Society of Mechanical Engineers. All rights reserved.