Mechanical engineers are helping to make the most complex machine of all-the human bodywork better and last longer, with projects ranging from safer skis to a new artificial heart. Researchers are gaining insights into designing a new breed of machines by using pneumatic artificial muscles to simulate the movements of a human arm. Current research is centered around alpine snow skiing, which produces complex loading at the knee. This load can be represented by three force and three moment components. Safety requires the bindings to release before loading reaches a level where it could injure the knee. The cornerstone of the artificial-heart project is the development of a novel magnetically levitated turbo-blood pump, which is intended to serve as a remedy for people with what would otherwise be terminal heart disease. Named the StreamLiner, its design evolution is being governed by numerical optimization algorithms, thus automating the design process for maximum functionality and biocompatibility.


In recent years, more and more engineers—particularly mechanical engineers—have come to devote their careers to the specialty known as bioengineering. Bioengineering research is almost always done by a cross-functional team comprising experts from various disciplines; the key elements of these research teams, however, are almost always mechanical engineers.

“Traditionally, we think of the human body strictly as the realm of physicians and God,” said Jeanette Nusemblatt, a mechanical engineer who works as a bioengineer at the University of California, Berkeley. “However, in many ways it is a machine just like man-made devices. As a result, the skills of a traditional mechanical engineer—including design, dynamics, solid mechanics, controls, fluid mechanics, heat and mass transfer, thermodynamics, robotics, and manufacturing—are often needed to develop technological solutions to medical problems.” Most biomedical research is based at colleges and universities, but a number of companies are also in this field, most of which are relatively small and new.


The range of projects falling under the bioengineering umbrella is diverse and far-reaching. Some of the principal fields include artificial intelligence, biological micro-sensor fabrication, manufacturing of surgical instruments, sports medicine and sports injury research.

The University of California at Berkeley has many programs in the bioengineering field, among them a project that focuses on making skis safer—a particularly timely project in vi&w of recent celebrity skiing deaths. The program started in the late 1960s. In the first experiments, researchers attempted to measure the pressure along a ski during normal skiing. The project continued long after these experiments, however, with the focus and testing equipment changing along the way. One basic credo remains the same: Obtaining data in the field from an actual skier is more effective and valuable than simulating skiing conditions in a laboratory or on a computer.

Current research is centered around alpine snow skiing, which produces complex loading at the knee. This load can be represented by three force and three moment components. Safety requires the bindings to release before loading reaches a level where it could injure the knee.

Two sets of data are critical to binding design: the relationship between moment components at the knee and forces and moments at the bindings, and the simultaneous limits of the three moment components generated at the knee during normal skiing. Bindings can be designed to release when the set of three moment components at the knee falls outside of the moment envelope. The forces and moments acting on the bindings are related to the three moment components at the knee through a predetermined model.

The project’s purpose is to determine the moment envelopes at the knee and the force and moment envelopes at the toe and heel bindings during normal skiing, and evaluate the characteristic space of the envelopes. Researchers are also determining the dependence of the moment envelopes at the knee and the force and moment envelopes at the bindings on individual characteristics, such as age, height, weight, leg strength, muscle fatigue time, skiing style, and skiing ability. The results are aimed at guiding the design of new bindings and their settings by developing a model to correlate the envelopes at the bindings and knee.

To accomplish this task, a skier wears state-of-the-art skis and skiing equipment. While the subjects ski aggressively, dynamometers measure the forces and moments at the toe and heel. Simultaneously, ankle flexion, knee flexion, and muscle activity of the rectus femoris and biceps femoris are measured. All these data are stored on a laptop computer in a backpack worn by the subject. Data are transferred to a ground station via a wireless local-area-network connection. Throughout the run, a video recording is also taken.

In many ways, the human body is a machine that works just like man-made devices.

The data are collected on the slopes during high-force and high-moment-generating skiing: ankle and knee flexion using potentiometers, forces and moments at the toe and heel of the left ski binding using two six-degree-of-freedom dynamometers, and muscle activity using mean rectified surface electromagnetic generators. Data reduction is performed after each run at the ground station. Analysis of the data is conducted in the university’s Dynamic Stability Laboratory.

The Most Complex Pump

The McGowan Center for Artificial Organ Development at the University of Pittsburgh is working on a new kind of artificial heart. This project, highly significant in its scope and potential benefits, requires skills from virtually every discipline of mechanical engineering, including machine and pump design, fluid flow, and heat transfer.

“Because of the exacting demands posed by chronic human application of artificial organs, we believe that their design requires a quantum paradigm shift from conventional methods,” said program director James Antaki. “Our laboratory is dedicated to creating new technology for the development of blood-wetted internal organs. This entails adapting the latest computer tools from aerospace, such as computational-fluid-dynamics analysis, computer-aided design, and rapid prototyping, to the design of medical devices.”

To bridge the gap in applying these tools to artificial organs, the lab is developing unique theoretical, computational, and experimental tools and techniques. Some of the active areas of research include constitutive modeling of blood, computer simulation of transport, advanced flow visualization, blood damage modeling, computerized shape optimization, and multidisciplinary design. To ensure that these systems are responsive to physiological demands and provide a satisfactory quality of life, the group is also developing models and methods of control and power generation.

The cornerstone of the artificial-heart project is the development of a novel magnetically levitated turbo—blood pump, which is intended to serve as a remedy for people with what would otherwise be terminal heart disease. Named the Streamliner, its design evolution is being governed by numerical optimization algorithms, thus automating the design process for maximum functionality and biocompatibility. Four such pumps are currently in development: two magnetically levitated, computer-optimized blood pumps, the Streamliner Zephyr and the Streamliner Hemoglide; a miniature centrifugal blood pump with biocompatible bearings; and a mixed-flow pump that includes the added feature of gas transport.

The StreamLiner device is intended as long-term-to-permanent therapy for patients suffering from end-stage heart failure. The device would be located in the abdomen and connected to the native heart by a pair of tubes. Implanted with the pump would be an internal control circuit and battery. Outside the body, the patient would wear an external battery pack and controller, which would couple to the implanted circuitry.

Unlike the artificial hearts now in use, which were designed with the intention of imitating nature, the StreamLiner does not resemble or function like a natural heart. Rather than using flexing diaphragms and prosthetic heart valves, the StreamLiner uses a small propeller to create blood flow. To eliminate the risk of failure due to bearing wear, the impeller of the StreamLiner is suspended by magnetic levitation similar to that used on magnetic-levitation passenger trains.

The design features of the StreamLiner are intended to avoid many of the problems associated with fatigue and blood clotting and will presumably lead to a more reliable, less complicated, and less expensive device for physicians to offer their patients. Other organizations—including NASA, Carnegie Mellon University in Pittsburgh; General Electric Corp. in Schenectady, N.Y.; and the U.S. Department of Energy in Washington, D.C.—are working with the lab on various phases of the project.

A novel aspect of this research is the concept of coupling computer optimization with fluid dynamics analysis. This combination of technologies will provide designers of artificial organs with a powerful tool for rapidly developing devices with superior fluid dynamics and transport characteristics, which are more efficient and cost-effective.

Although the design of the StreamLiner pump is the crux of the project, there are many other aspects to the design of the system. For example, a separate team is developing improved models for describing and predicting blood flow behavior, both in natural blood vessels and in artificial devices. This research is also aimed at understanding the implications of artificial circulatory support for damage to blood cells and alteration of the physical properties of blood.


The shear stress at the interface between flowing blood and its boundary is also of extreme importance in the design of biomedical devices. Current techniques for measuring shear stress at the blood/biomaterial boundary are limited in their precision and resolution. The most precise techniques provide measurements only at discrete points along the wall, while the more global measurements rely on indirect calculation of shear from discrete velocity measurements. For unsteady flows, there is no easy technique that provides reasonable temporal resolution. Researchers are now working on developing a photochromic paint that, when applied to the interior surface of a transparent replica of the device or vessel, will alter its color in response to shear stress.

Another important facet of the project is the control system. With a heart, this is not just a matter of setting a fixed flow rate. Rather, the device must respond to the demands of the body and react to changes in activity, posture, and stresses; this requires specialized algorithms and circuitry. The center’s ongoing program to develop a hemodynamic controller for rotary blood pumps is now concentrating on three key areas: ensuring robustness while maximizing performance, ensuring reliability of sensorless state estimation, and adapting the controller to variations in the demand set point.

The controller team is pursuing a hybrid architecture, which fuses heuristic rules for detecting discrete events such as atrial suction, continuous control for regulating the hemodynamics, and adaptive, performance-based control for updating the hemodynamic set points according to demand.

So far, preliminary results are satisfactory. However, the accuracy of the estimator may be adversely affected by dramatic changes in blood-flow parameters or transient noise. As a result, a system supervisor has been introduced to monitor the status of the state estimator and to detect fault conditions. Based on estimates of operating range, it selects between performance-based, heuristic, or fail-safe modes of control.

Development of all phases of the artificial-heart project is ongoing. Test trials on animals are expected to begin within a year.

Imitating Life

Most bioengineering research is dedicated to making the human body healthier and safer. However, the Bio-robotics Laboratory at the University of Washington in Seattle has turned that around. Rather than developing machines to help people, it is using human anatomy as the theoretical basis for the design of a new breed of machines.

The lab’s Anthroform Arm Project is developing a mechanical replica of the human arm for the purpose of studying the biomechanical and low-level neural properties of the arm and spinal-reflexive postural control. The arm is constructed from a variety of subsystems that emulate the corresponding structures of the human arm and central nervous system. Skeletal components are constructed with fiberglass, using molds made with human bones extracted from cadavers. Stainless-steel surgical replacement joints are used for the elbow and shoulder joints. Fabric ligaments are attached to both sides of the joints at biomechanically accurate locations.

Pneumatic artificial-muscle actuators, which are attached at biomechanically accurate locations, are used for the arm’s muscles. These devices consist of an expandable internal bladder (an elastic tube) surrounded by a braided shell. When the internal bladder is pressurized, it expands like a balloon against the braided shell. The braided shell acts to constrain the expansion to maintain a cylindrical shape. As the volume of the internal bladder increases due to the increase in pressure, the actuator shortens and produces tension if it is coupled to a mechanical load. The force generated by the artificial muscle is dependent on the weave characteristics of the braided shell, the material properties of the elastic tube, the actuation pressure, and the muscle’s length.

Functional replicas of muscle spindles were developed to sense the elbow position. Like human muscle spindles, these sensors contain active elements that change their output functions based on separate neural input signals.

New computational models of the human neural circuits used in spinal reflexive postural control were developed from known physiological parameters and existing experimental data. A specialized digital-signal-processor board has been developed to emulate human spinal-reflexive neural circuits used in postural and motion control.

“The basis for this project is that nature doesn’t make mistakes,” said principal researcher Joseph Alicandri. “The human body is more versatile and complex than any machine ever designed by a human. We are striving to adapt the best elements of our own designs into the design of machines.”