LightProbe - Thermal-Power aware on-head Beamforming
In the LightProbe project, we are exploring the next generation of medical ultrasound imaging systems: The LightProbe is a programmable ultrasound transducer head, which incorporates the entire analog frontend and directly outputs the captured digital samples. This allows the LightProbe to be directly connected to any commodity hardware (phone, tablet, workstation) for post-processing over a standard digital link as simple as a standard peripheral, like a camera. The LightProbe has two main challenges: The involved ten's of Gb/s data-rates produced by the frontend, which need to be processed and transported off-head, and the power/thermal constraints of such a high-performance handheld device.
The goal of this project is to design and implement a thermal-power aware ultrasound beamformer to be placed on the FPGA situated in the LightProbe ultrasound head. The beamformer is the core processing unit in any ultrasound imaging system as it produces the image from the raw sensor data. Similar to other handheld devices like smartphones, the LightProbe is a thermal power limited device as no active cooling elements can be easily integrated. Continuous operation on maximal performance may heats up the device surface beyond 43C allowed for medical devices in contact with human skin. Thus it must be able to throttle it's performance when not needed, e.g. , a 30 fps image is sufficient to find the right placement of the probe for the examination. Once the placement is found, a high performance mode can be activated for a limited period of time providing multi-khz imaging to extract tissue feature like stiffness or vector blood flow measurement.
The goal of this thesis is to design and implement an online reconfigurable beamformer than can adapt to different performance modes and a thermal power aware control unit, which configures the beamformer based on real-time temperature sensor data from the system.
- Looking for Interested Students
- Supervision: Pascal Hager
- 50% Beamformer design & implementation
- 20% Thermal characterization & modeling
- 30% Implementation and Test of the thermal power aware controller
- VLSI I, C, Matlab