Parkinsonian tremor measurement using Radar technology
3D mapping of Parkinsonian hand tremor using advanced radar technology [Walabot] for accurate muscle motor fluctuation sensing.
Worldwide 10 million people are affected with Parkinson’s Disease(PD). 1% of population above 60 has PD, also 10% of all patients develop symptoms before age 50. Unfortunately PD doesn't have cure,alternatively there are treatments to reduce the effects of PD. Any treatment needs a quality feedback system to improve its methods, and now technological advancements has enabled us to gather more data with high precision. Parkinsonian tremor is one of the effects of PD which can be quantitatively assessed. Parkinson’s tremor associates to muscle motor fluctuations and dyskinesia, so data measured for the tremors can help in better understanding and also improve treatment methods. To measure hand tremors different approaches have been made, some of them include wearable IMU(Inertial measurement Unit), touch sensitive apps and other electromagnetic motion tracking methods. Due to the fact that non of these methods 3D map the tremors, I am trying a new approach with an advanced technology to get detailed data.
What am I trying to solve ?
Parkinson's tremor assessment has two parameters 1) amplitude 2) frequency of the shake, these two are used to studying and monitoring the severity of the tremor. Consider the wearable device method, where an IMU is attached to patients hand and data is gathered. Since IMU’s is attached to a single point, it can measure acceleration and gyro data in 3 directions at that point. But problem is that tremors is measured from a single point doesn't enrich us with full details about the tremors. If the tremors can be characterised using data from multiple points simultaneously or similar to a 3D mapping it can significantly improve the characteristics of the tremors. The quest to find an device to solve this issue is plotted here.
What have I found ?
There are several 3D mapping technologies available, Radar, Lidar, IR and camera. Each method has its own advantages and disadvantages, but to looking for low cost and easy to develop method is quite hard. Uniquely, due to the advancements in technology, systems have shrunk and prices have gone down considerable.
Walabot is one advanced 3D radar technology, that has been shrunk into a small package. It can map objects, track motion in a way it is quite hard to measure using traditional instruments. So Walabot is a potential candidate to solve this problem, and it will be experimented in detail.
Exploring and Experimenting
What is Walabot ?
Walabot is small device which uses radio frequency to sense the environment around it and detect objects. It has array of linear polarised broadband of antenna from which it transmits, receives and records RF signals. Recorded data is processed and sent to computer or smartphone for further processing. Using the data, it can: do in-Room Imaging, detect objects, locate and track objects, motion Sensing (i.e. Breathing Patterns, Gestures), speed measurement, in-wall imaging.
What am I measuring?
Primary objective of this project is to detect the shiver in the hand. Walabot demo application gives idea for our proof of concept, that it can measure object in 3D and it can also measure its movement.
Using Walabot demo application I configured it to detect objects within 50cm with 1mm of accuracy. It is clearly seen that it can measure the amplitude and frequency of the shiver. In conclusion, the demo application shows that Walabot can gather data for hand tremors in 3D as shown in picture.
Since I have the proof of concept is ready, next step is to develop my own code to record and process the data.
Application Development
Walabot provides SDK to code and develop applications. I am using Walabot SDK with Matlab to get data from device and find the amplitude and frequency of change. Check the code below.
I am performing two types of experiment to evaluate my code.
- High movement with low shiver (i.e High amplitude low frequency)
- Medium Movement and high shiver (i.e Medium Amplitude ad high frequency)
Interpreting the Data
Output signal from Walabot is processed and plotted on graph. It is clearly seen that the frequency and the amplitude of the tremor are different and it can be measured. Further coding can be done to get dominant frequency and amplitude in much more detail.
Future work
It is clear that using the data from Walabot, Parkinsonian tremors can be quantified in more details. As future development more experiments can be done to check different stages for PD development.Similarly, in any treatment early detection can significantly improve treatment process as well as it can help caution future patients. So Walabot 3D mapping can be used for early detection. There has been tremendous improvement in machine learning algorithms, so our 3D mapping data can be utilised with machine learning for early detection. Machine learning needs huge data sets, so that it self is project of its own. So will leave that to future.
Conclusion
It started as a fun project if Walabot can detect tiny movements, after understanding the potential of the device it has opened lots of opportunities. Number of PD patients has been increasing since last decade, also the number keeps increasing. This miniature Radar technology can be used in medical field to improve the treatment process and used in early detection as well.
