Dynamic Interrogation using Bimodal Sensing and Statistical Game Control

Many of the current medical imaging devices on the market are static in the sense that the source and detector positions are fixed. If the source and detector positions change intelligently, superior performance medical imaging devices could be created. Developing such a high-performance system demands novel sensor and actuator coordination.

We will design, build, and test a Dynamic Interrogation System that will be able to differentiate between benign and malignant tumors. The system will non-invasively characterize tumors by dynamically adjusting the sensor geometry. Our system will capture the elasticity and physiological changes of the lesions when pressure is applied from the surface via robotic arms.

A novel cooperative control theory will be developed for the robot arm control. This proposal displays both scientific and conceptual innovation as it works to develop a new sensing system that provides the touch and color properties of a tumor through dynamically obtained images. This will be especially useful in medical robotic systems because it would allow for more accurate performance, deeper interrogation depth, and larger interrogation areas.

The Dynamic Interrogation System that identifies malignant tumors in a non-invasive and harmless (no ionizing radiation) manner will have a significant impact on the screening, diagnosis, and biopsy rate. One particular application for this medical imaging device is in reducing breast tumor overdiagnosis, which would lead to improved mortality rates and overall reduced health care costs.

Perhaps, the main benefit to society may be that the device will accurately assess the risk of breast cancer for women in rural and remote regions. We will train one graduate student and two undergraduate researchers in this project and will also introduce a new Sensing Systems course for engineering students. Our prototype system will be the focus of outreach activities for high school students through Engineering Open Houses, and K-8 children through the Ayuda Community Center Summer Camp, which serves low-income, African-American and Hispanic children in the north Philadelphia area.

This project's goals are to (1) develop a dynamic interrogation system, (2) intelligently coordinate the sensor/actuator geometry using Stackelberg Statistical Game Control, and (3) integrate the system and test the dynamic interrogation system. Conventional X-ray source/detector geometry is static, so it leads to two-dimensional information. Optimally varying the source and detector geometry will lead to more accurate three-dimensional information.

The scope of this project is to develop a procedural game control method to improve the performance by dynamically changing the spatial geometry of the sensors. We will develop a Dynamic Interrogation system, where the position and orientation of the light source and the position of the detectors are optimally controlled to characterize embedded inclusions.

Viscoelastic and physiologic properties of tumors will be measured by Tactile Imaging Sensor and Diffuse Optical Spectroscopy, respectively. These bimodal sensors will be dynamically controlled with the robotic manipulators. For dynamic sensor/actuator geometry control, a novel game control strategy, Stackelberg Statistical Game Control, will be developed.

This is a leader-follower type of game control strategy that shapes the cost distribution. Consequently, the position and orientation of the sources and sensors will be optimized. This will allow the system to measure tactile and spectral properties with higher accuracy.

We propose to develop bimodal breast tumor models for testing. Investigators will determine the sensitivity and specificity of detecting malignant tumors using one hundred chicken breasts. The system will non-invasively characterize inclusions by dynamically adjusting the sensor/actuator geometry through a novel control theory.

Finally, dynamic sensor/actuator control will allow the system to have higher sensitivity and specificity compared to a static system. This project will advance the field of sensing systems by intelligently controlling sensor geometry and integrating multiple modalities.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.