Home > Research > Haptic Interfaces and Virtual Environments
Haptic Interfaces and Virtual Environments
Needle insertion is a challenging task, especially in areas with little room for error. Brachy-therapy for prostate cancer, for example, is the localized placement of sealed radioactive sources (seeds) using 6-inch long, 18-gauge needles. The target area for seed delivery is a small volume (40-50 ccs) with limited access and through a small cross-sectional area of the perineum. At least 20 needles must be inserted to place anywhere from 80 to 150 seeds. These constraints make the procedure very difficult to perform. To practice, physicians use “phantoms,” or non-biological substitutes for human flesh and bone. Even the best phantoms, however, cannot come close to the complexity of textures and the individual variability of true human anatomy. Alternatively, some surgeons are turning to virtual environment training tools to practice their technique. One of these, which trains surgeons for endoscopic sinus surgery, uses a virtual environment created from MRI images. But because these images are limited to data about shape, not stiffness or texture, doctors are not satisfied with it, insisting that it does not “feel” realistic.
Professor Allison Okamura envisions another approach, one in which a robot plays an important role. 
Okamura is an expert in robotic haptic exploration. When we touch something, we have a “haptic” experience of that object that includes sensations such as texture, hardness/softness, weight, temperature, and shape. Scientists know very little of how the brain processes sensory experiences, particularly haptic ones, and so it is impossible at this point to recreate a true haptic experience in virtual reality. But Prof. Okamura is tackling the problem one step at a time, using simple robotic systems and complex models. Eventually, she hopes to help build a virtual reality environment that can give feedback on many of the important pieces of a haptic experience, making virtual reality “feel real.”
The model she is developing must translate data that a robot takes from a real environment into a motor-driven combination of acceleration, force, and position that will be delivered to a human hand on the other end of a stylus, thereby allowing the human to “feel” the environment that the robot does. Prof. Okamura’s masters’ student, Christina Simone, is currently working on a needle insertion system in which a robot performs many real needle insertions in various organs, all the while recording data on force, acceleration, and position.
Once the haptic experience of poking is fully modeled, she can move her robot on to using scissor-like tools, exploring the forces involved in grasping and cutting. Eventually, these individual haptic experiences can be woven together to form a modeled “reality” that feels more like the real thing.
The data collected in Okamura’s research could find use in various applications such as training simulations, telerobotics, and cooperative (assistive) robots. In robotic procedures, data coming in real time could be compared to some idealized model and used to give the doctor feedback on his/her performance. In addition, feedback from real needle insertions could also be compared with data collected from phantoms and used to improve the materials that make up phantoms, making them more realistic. The possibility exists, notes Okamura, for autonomous robotic procedures, where the robot does an entire procedure on its own, although she believes that these are many years away.