Biotriboliogy Gallery


AMTI’s VIVO™ brings new life to joint motion simulation, dramatically increasing the kinetic and kinematic realism of simulation to provide the closest possible approach to in-vivo conditions. VIVO™ provides the world’s first full-speed, full-load six degree of freedom environment in which to test joint prostheses as well as biological joint specimens.
VIVO™’s performance highlights include:
  • Six degrees of freedom – all joint motions can be controlled and measured
  • Each axis may operate in force or position control – including zero-force control to simulate an unconstrained axis
  • Convenient user programming of force and motion profiles
  • AMTI’s exclusive multi-fiber ligament model – captures 6-DOF soft tissue inter-axis coupling
  • Traditional heuristic soft tissue model also supported
  • Grood and Suntay coordinate system – kinematics are described using joint-referenced virtualized axes
  • User-defined coordinates – change coordinate system origin and orientation to accommodate test specimens as mounted
  • Iterative learning control – advanced feed-forward control system provides stability and rapid convergence to command during cyclic testing
  • Temperature-controlled serum circulation and containment system
  • One to three independently programmable test stations in a single frame – reduces capital acquisition cost, floor space requirement and power connections
  • Optional VivoSim visualization software – provides an accurately scaled real-time display of the Vivo machine, the component under test, and the multi-fiber ligament model


Degree of Freedom / Axis Specification
Axial load ±4500N
ML / AP load ±1000N
Moment capability ±80 N-m FE
±60 N-m FE
±50 N-m FE
Flexion/Extension 200° (Cartesian)
110° (Grood & Suntay)
Internal/External ROM ±40 degrees
Ab/Adduction ROM ±25 degrees
vertical translation 50 mm stroke
ML / AP translation circular workspace
50 mm diameter
*specifications subject to change without notice


VIVO™ provides accurate robotic joint motion simulation for knee, hip, shoulder, temporomandibular, elbow, ankle and spinal joint motions. VIVO™ may be programmed with standard or custom force and motion profiles, supporting research and test protocols that range from standard activities of daily living (ADL) tests through extreme sports and even simulation of accidental conditions and injurious activities.

VIVO™ supports two virtual soft tissue models. VIVO™ introduces AMTI’s exclusive, patent-pending multi-fiber ligament model, which can capture the 6-DOF inter-axis coupling typical of biological soft tissue systems. This major advance in soft tissue modeling capabilities opens new doors in research and equips VIVO™ to handle anticipated future test specifications. VIVO™ also offers our patented heuristic constraint model. Upward-compatible with the soft tissue model used in our older simulators, in VIVO™ the heuristic constraint model has been enhanced and extended with greater flexibility and programmability. An independently-specified heuristic constraint model is available for every axis of the VIVO™ that is in force-control mode.

VIVO™ is the world’s first simulator to fully support the Grood and Suntay joint coordinate system, accepted by ASTM, ISO and ISB as a standard for biomechanics joint testing. The user commands forces and motions with respect to software-defined joint-referenced axes, making it unnecessary to consider the complex, pose-dependent, compound-rotation kinematics of the machine’s physical actuators. The virtual coordinate system is also the key enabling element in VIVO™’s exclusive new test setup system. VIVO™ can sense how the joint or implant is placed in the workspace and relocate the coordinate origin in software to match the installed pose of the joint. This capability accommodates mounting variations in the experimental setup. It relaxes the level of accuracy required when mounting joint specimens in the adapters that hold them in the machine.

VIVO™ has a new, patent-pending iterative learning control system to improve the accuracy of the actual forces and positions generated by the simulator. Based on modern control theory, this feed-forward, self-tuning frequency-domain implementation requires less user configuration than ever before, yet provides the best stability and fastest convergence available in any AMTI simulator.

Capable of performing short-term kinematic and long-term durability evaluations, VIVO™ is a configurable system that may have from one to three joint test stations in a single frame. The stations operate independently of each other. The VIVO™ user interface runs on a separate Windows computer host. The UI software provides one-click copying of setup and programming between stations for situations where the same test protocol is to be run on multiple stations.

VivoSim is an optional software product that aids understanding by displaying an accurate 3-D model of the joint components and the multi-fiber ligament model as they move in space. While VIVO™ by itself is a completely functional, stand-alone simulation system, VivoSim provides an enhanced ability to look inside the multi-fiber ligament model and examine the strain, tension, and resolved force components individually for each fiber. VivoSim also has a near-real-time stand-alone modeling capability. VivoSim, sold separately, is described elsewhere on the website.

VIVO™’s unique combination of speed, range of motion and force capability, programmability and virtual soft tissue models enables testing of real-world implant failure modes, such as adverse edge loading conditions, micro-separation, stem and cup impingement, condylar liftoff, and joint subluxation.

VIVO™ is the most realistic simulation system available.


VIVO™’s workspace and compact packaging are designed to facilitate human joint studies while minimizing the use of laboratory floor space.

VIVO™ system consists of one to three test stations assembled and shipped as a unit. Each station is equipped with six servo-hydraulic actuators. Acquisition and installation costs are optimized by sharing a single electrical power connection, realtime controller, hydraulic pressure supply, and hydraulic return for all stations in a frame. Although there is a single realtime controller, it executes independent control loops for each of the stations. Therefore the stations are programmed and operate independently.

The unique actuator configuration on the lower stage provides a floating instant center of rotation. In combination with the software-defined virtual axes of the Grood and Suntay coordinate system, many of the joint alignment issues found in legacy test machine designs are eliminated.

Precision displacement sensors are co-located with the hydraulic actuators to generate position feedback for the control system. Each station has a six-axis force sensor, which measures the contact forces and moments for force feedback. The lower part of the tested joint is mounted directly to the force sensor, achieving close coupling between joint contact interactions and the feedback sensor. Force disturbances arising from actuator nonlinearities or imperfections are included in the force feedback measurement and become correctable by the control system. If desired, the mass and polar moment of inertia coupled to the force sensor may be entered so that the control system can cancel the effect of inertial body forces.

Each station has a temperature-controlled serum containment and circulation system for tests that are conducted in a fluid environment. The thermal plate can heat or cool the serum to achieve setpoints between approximately 10°C and 45°C.

The stations in a multi-station VIVO™ frame operate independently. However, the VivoControl UI supports one-step copying of programs and setups between stations so that the same test protocol can be executed for multiple samples.


AMTI's extensive biomechanical simulation experience coupled with modern advances in control technology has culminated in the new VIVO™ control system. It is the most sophisticated robotic control system available today for joint motion simulation. The control system provides two kinematic modes.

Joint Coordinate System mode - implements the Grood and Suntay Joint Coordinate System (JCS). The Grood and Suntay joint coordinate system has been adopted by the International Society for Biomechanics, ASTM and ISO. In G&S mode, control inputs and data outputs are resolved along joint-referenced axes that coincide with clinically-meaningful directions – medial / lateral, posterior / anterior and distraction / compression translations, and flexion / extension, abduction / adduction, and internal / external rotations. The mapping function between actuator positions and Grood and Suntay coordinates is computed from a reference pose setup – the user identifies the Grood and Suntay coordinates of a defined joint pose, then produces that pose on the test sample installed in the machine, and selects command on the UI. There is also a pre-defined default mapping, which may be selected at any time. Once the kinematic mapping is defined, the control system updates the relationship between the physical actuator positions and Grood and Suntay coordinates 2000 times per second. This operation assures that the Grood and Suntay axes maintain their joint-referenced definitions for all machine poses within the physical workspace of the VIVO™. In G&S mode the flexion / extension axis has a range of motion of 110°. Using VIVO™’s setup features, this physical range of motion can be associated to any 110° window of the virtual G&S flexion coordinate, subject to limits of ±180° on the coordinate value. In G&S mode, every axis may operate in position-command or force-command mode. The command mode is independently selected for each axis and any combination is possible.

Cartesian Coordinate System mode - for compatibility with traditional machines. In Cartesian Coordinates mode, input and output translations and linear forces are resolved along an orthogonal X-Y-Z coordinate system that is fixed with respect to the frame of the machine. Input and output rotations are resolved along rotational axes that coincide with the physical actuator axes of the flexion and ab/adduction actuators, and a virtual Z-rotation actuator. In Cartesian Coordinates mode the flexion arm has up to 200° range of travel. In Cartesian Coordinates mode the four axes of the lower stage can operate in force- or position-command mode. The flexion and ab/adduction axes operate in position mode only.

Command waveforms are generated by independent 1024-point waveform buffers for each axis. The waveform is interpreted as a position (translation or rotation) or force (linear force or moment) command according to the current axis command mode. Switching an axis between position and force command mode is as simple as ticking a box in the setup configuration dialogue. The speed of the waveform is controlled by setting the buffer period, which may be between 0.5 and 100 seconds (2 to 0.01 Hz).


VIVO™ introduces an entirely new version of AMTI’s iterative learning control (ILC) algorithm. This newly-developed, patent-pending system is implemented partly on the VIVO™ realtime controller and partly in the VivoControl host software. It advances the state of the art in stability, speed of convergence, residual error and ease of tuning compared with earlier versions of ILC.

The ILC system collects error data over an entire period of the programmed waveform. The error is transformed into an equivalent frequency-domain representation, and various processing steps are applied, including truncation of frequencies outside the range of interest, and inverse phase and magnitude compensation for the axis transfer functions. The result is converted back to the time domain and applied as an increment to the axis positions recorded on the previous cycle. Because of the batch-wise processing and cyclic operation of the waveform, this approach produces a feed-forward compensation that, in theory, is capable of driving the error to exactly zero over time. While non-repetitive disturbances in any real system will prevent true zero error, in practical applications the new system usually reduces error to well below 1% of command.

The learned compensation is automatically saved and can be used as the starting point after a test interruption. This capability is useful when a test is stopped temporarily for weighing, serum change, containment bag replacement, etc.

The new system includes a pre-run phase we call “haptic mapping.” Before starting a new test, a few cycles of the waveform are run at greatly reduced frequency, usually 1/10 to 1/20 of the desired full-speed. This reduced speed allows the basic P+I control system to operate with relatively low error, in effect measuring the pose-dependent compliance of the joint. Processing to account for the estimated axis dynamics at full speed generates a first-pass compensation. While this approach will not usually cancel the error completely, it can reduce errors during the initial cycles of a test by 50% or more, offering more rapid convergence and reduced chance of specimen damage during the early cycles of a test.

The ILC system can be disabled for low-speed or non-repetitive testing. In these cases, VIVO™’s P+I axis controllers provide the performance of a standard servo-hydraulic control system.


The implanted joint is a composite of natural biological structure and synthetic engineered structure. An accurate simulation of the kinematics, kinetics and durability of the joint structure requires accurate re-creation of the joint contact forces experienced in-vivo. Since soft-tissue forces can contribute very significantly to the joint contact force, a realistic simulation environment must include a way to simulate the effects of soft tissue forces.

VIVO™ provides two virtual soft tissue models.

The new multi-fiber ligament model is the result of a ground-up re-thinking of how to structure a kinematically- and kinetically-accurate virtual force model. It is capable of representing 6-DOF cross-coupling between joint displacements and the resulting modeled forces.

The heuristic constraint model in VIVO™ is an enhanced version of the constraint model used in AMTI’s previous simulators, now independently programmable for any axis in force-control mode, and able to represent 2-axis cross-coupling using any axis as the secondary axis.

The inputs to the constraint model depend on the selected model.

In the heuristic constraint soft tissue model, each axis may operate with either a single-input or two-input constraint table. To define a one-input model, 15 axis coordinate values are defined as interpolation points and the constraint force (or moment) at each interpolation point is entered into a table. During operation, the controller uses the instantaneous axis coordinate to interpolate into the table and calculate the constraint force. The two-input mode can capture cross-coupling between the main axis and any one of the other five axes. Eight interpolation points are defined for the secondary axis. The constraint force (moment) values are represented as 8 curves, each curve corresponding to one of the interpolation points defined for the secondary axis, and comprising 15 points of variation along the primary axis. The ensemble of 8 curves defines a surface that expresses the constraint force as a 2-parameter function of the main and secondary axis coordinates. During operation, the controller uses the main and secondary axis coordinates to perform a 2-D interpolation into the constraint table. Every axis offers a completely independent implementation of this model. The VivoControl UI offers tools to assist in generating the constraint force tables, such as linear, quadratic and polynomial-fit spring models. Alternatively, the constraint force tables may be generated in a prescribed CSV format by an external application and imported into VivoControl. VivoControl also provides CSV export of the constraint force tables. Among other uses, the heuristic soft tissue constraint model is typically required for tests that conform to today's ISO and ASTM standards. The heuristic soft tissue model is available in either the Grood and Suntay or Cartesian kinematic modes.

The multi-fiber ligament model arises from a re-imagined approach to soft tissue modeling. Rather than determining ahead of time what constraint force corresponds to every coordinate for each axis, in the multi-fiber ligament model, up to 100 ligament fibers are defined by specifying the (x, y, z) coordinates of their proximal and distal insertion sites on the joint components, their stiffness, and the amount of strain or slack the ligament has when the joint is in the defined reference pose. The length at the reference pose is automatically computed from the coordinates; combined with the supplied strain at reference pose, the zero-force length of the ligament is automatically computed. During operation, the controller calculates in real time the length and line of action of the ligament fiber. For engaged fibers, i.e. those stretched beyond their zero-force length, the force thus computed is applied to the Grood and Suntay coordinate axes. Moment-of-force is also computed for each fiber, based on the moment arms between the line of action and the Grood and Suntay axes. The contributing forces and moments of all fibers are summed and applied to the joint. These calculations are performed at every servo tick on the realtime controller. Therefore, they are inherently synchronized with the instantaneous kinematic configuration of the joint. VivoControl also provides CSV-based import/export of the multi-fiber ligament model. Customers may wish to use an external modeling tool to generate a multi-fiber ligament model. By exporting the model from the external application to the defined CSV-based format, then importing that file into VivoControl, tedious and error-prone manual entry of the model is avoided.

The multi-fiber ligament model enables the development of testing for sensitivity to ligament balance and post-surgical ligament condition while providing the complex inter-axis coupling exhibited by the natural knee. The concepts inspiring the multi-fiber ligament model may be used for future ASTM and ISO testing standards.


For compactness and maximum service life, VIVO™ utilizes an all hydraulic actuator design. The main bearing of the system is hydrostatically supported to provide low friction, more accurate control, and long life. AMTI's unique hydrodynamic-seal actuators reduce maintenance downtime and are designed to last for hundreds of millions of cycles.

separate hydraulic power unit is required for operation. AMTI can provide hydraulic power units sized for one-, two- or three-station VIVO™ systems. Customers can use their own infrastructure if preferred. Please consult with AMTI regarding requirements.


VIVO™ sets a new standard of protecting your valuable test investment. Every physical quantity is continuously monitored. In the event of a fault or an overload, the controller responds to protect the machine and the sample. User-defined limits and a programmed action can also be created for many dynamic variables. This rapid response prevents inadvertent specimen damage during setup as well as during operation of the machine.

Operator safety is enhanced with a light curtain safety interlock that encloses the working volume of the machine. The light curtain may be disabled with a key switch for maintenance or setup operations that require the machine to be energized. Disabling the light curtain protection requires enabling two overrides: the key must be inserted into the key switch and the switch turned to the clearly-labeled “DISABLE” position, and an override setting must be selected in the VivoControl UI. Turning the key switch to the “DISABLE” position also activates a flashing alert light. The key cannot be removed from the key switch when the switch is in the “DISABLE” position. Therefore, a site protocol regulating access to the key can be used to control the operator’s ability to defeat the light curtain protection.


The VivoControl UI software runs on a Windows PC in a compact desktop box, with several uncommitted USB ports, an optical drive, and two gigabit Ethernet adapters. VivoControl is pre-installed by AMTI. One Ethernet adapter is configured by AMTI and is reserved for VivoControl to use in communicating with the VIVO™ realtime control computer. The other Ethernet adapter is available for connection to the customer’s LAN. Customers can also install a WiFi adapter if desired. The computer, keyboard, mouse and monitor are crated and shipped with the VIVO™.

The VIVO™ is designed to run as a stand-alone system and all VivoControl features relating to the management, control and programming of the VIVO™ are available without a LAN connection. Users can connect the host computer to a LAN to facilitate transferring files on and off the VIVO™ host computer for post processing of results, archiving setups, transfer to other VIVO™ installations, etc. Since the host computer is a standard Windows client, it integrates readily with the IT infrastructure at most sites.† If there is no LAN connection then file transfers will require physical media.

Integration with the VivoSim real-time visualization software is an optional feature that requires a LAN connection. (VivoSim is sold separately and is described elsewhere.)

†AMTI ships the host computer with Windows updates disabled, and strongly recommends this configuration, to prevent the chance that an update could interrupt a long-running test. Corporate IT policy might discourage or disallow a LAN connection for a client with updates disabled.


VIVO™ is capable of performing tests to ISO 14242-1, ISO 14243-1, ISO/CD 14243-3, ISO 14879-1, ISO 16402, ISO 18192-1, ISO/TR 22676, ISO 7206-4, ASTM F1223-08, ASTM F2790-10, ASTM F2694-07, ASTM F2777-10, ASTM F2028-08, and ASTM F1829-98.


To learn more, contact us at

© Advanced Mechanical Technology, Inc.
176 Waltham Street, Watertown, MA 02472-4800 USA
1-617-926-6700 | Contact Us