Research Areas
Energy Systems
Overview
Integrated approach to energy systems is crucial for enabling various energy sources, efficient energy usage, electricity pricing and data-driven assest management and planning. The current challenge is meeting the growing demand of electricity with smart grid operations.
At Signal Science Lab, we are developing advanced data structures for comprehensive and integrated analytics of smart grid, as well as mobile and cloud computing algorithms for smart grid and reliable pricing. Our expertise lie in processing large data sets in real-time, while maintaining the data privacy. We combine AI models with mathematical approaches to yield deterministic results.
From load forcasting to business inteligence dashboard, the outcomes of our research allows utilities for improve their operations and have access to early warning systems, as well as improved supply chains and customers paying lower prices.
Products
Predict-a-Bill Software
- Subscription-based billing system that converts variable energy costs into fixed monthly pricing.
- Reduces billing related service calls volume.
- Reduces customer switching cost when a resident moves out and moves in.
- Preempt defaulters in billing months in advance.
Load4Cast Software
- Substation load forecasting for asset maintenance and planning.
- Real-time monitoring and access-analytics of substation.
Grid IQ Software
- Grid intelligence for cross functional data infrastructure for accessing key performance index (KPI).
- Mapping of consumer-level AMI data to substation and feeder level analytics.
The Story of New Ideas
It will never work
It will work, but only for short time
The guys are really lucky
They are right
New Ideas
Unsolved Problems
Improving the Rate of Noise Reduction to N/2
Signal averaging is widely used method to reduce noise. However, it reduces the noise via sqrt of the number of scans (N), which is slow by many standards. Can there be a method that can reduce it by a linear rate, say N/2. There needs to be a mathematical framework that accomplished the goal for various types of random noises. The method can be first tested on integers before applied on the real numbers.
Solving Sudoku via Mathematical Equations
Sudoku is a popular game with the ability to improve brain efficiency. While everyone has there own strategy, it is possible for to solve Sudoku problems via mathemational equations. The rules for the game is well defined and hence can be converted into a set of equations. The question is how to solve it via linear algebra.
Prime Numbers and Number Systems
It is well known that the prime numbers are only divisible by themselves and one. The questions that is known to us (the lab members) if it is also true is the number system is changed. For example, a prime number will function as it is in a number system (base 3 to 10), but if we convert a prime number from base 10 (say 7) to base 4, will it retain as a prime number. More importantly, if possible, is this property universal, i.e. for all prime number at all number systems.
Simultaneous Image Compression of Multiple Images and Videos
Digital images are now pervasive and the most important activtity involved with is sharing via app, such as WhatsApp and WeChat. One may have noticed the poor quality or the reduced image size. Sharing images cost bandwidth and so does storing them. Can one have an compression allgorithm that compress all the image at once, without compressing them individually, in a lossless manner. For that, one would need a entropy conserving binarization algorithm which can convert decimal number system into binary format without reauiring additional storage space. The mathematical proof of such algorithm needs to be carried out first.