Bharath Sudharsan

Bharath Sudharsan, Dineshkumar Sundaram, Pankesh Patel et al.

The majority of IoT devices like smartwatches, smart plugs, HVAC controllers, etc., are powered by hardware with a constrained specification (low memory, clock speed and processor) which is insufficient to accommodate and execute large, high-quality models. On such resource-constrained devices, manufacturers still manage to provide attractive functionalities (to boost sales) by following the traditional approach of programming IoT devices/products to collect and transmit data (image, audio, sensor readings, etc.) to their cloud-based ML analytics platforms. For decades, this online approach has been facing issues such as compromised data streams, non-real-time analytics due to latency, bandwidth constraints, costly subscriptions, recent privacy issues raised by users and the GDPR guidelines, etc. In this paper, to enable ultra-fast and accurate AI-based offline analytics on resource-constrained IoT devices, we present an end-to-end multi-component model optimization sequence and open-source its implementation. Researchers and developers can use our optimization sequence to optimize high memory, computation demanding models in multiple aspects in order to produce small size, low latency, low-power consuming models that can comfortably fit and execute on resource-constrained hardware. The experimental results show that our optimization components can produce models that are; (i) 12.06 x times compressed; (ii) 0.13% to 0.27% more accurate; (iii) Orders of magnitude faster unit inference at 0.06 ms. Our optimization sequence is generic and can be applied to any state-of-the-art models trained for anomaly detection, predictive maintenance, robotics, voice recognition, and machine vision.

Bharath Sudharsan, Manigandan Chockalingam

Approximately 6.2% of the world's population (466 million people) suffer from disabling hearing impairment [1]. Hearing impairment impacts negatively on one's education, financial success [2][3], cognitive development in childhood [4], including increased risk of dementia in older adulthood [5]. Lack of or reduced social interaction due to hearing impairment affects creating or maintaining healthy relationships at home, school and work [5]. Hence, hearing impairment genuinely affects the overall quality of life and wellbeing. The cocktail party effect, which is a healthy hearing individual's ability to understand one voice in a cacophony of other voices or sounds, is an important ability lacking in people with hearing impairment. This inability results in difficulties with simple daily activities such as partaking in group discussions or conversing in noisy restaurants [6]. This smart hearing aid aims to provide much-needed assistance with understanding speech in noisy environments. For example, if a person wants to partake in a group discussion, he/she needs to place the microphone array based unit on a flat surface in front of him/her, such as a table. When conversations take place, the microphone array will capture and process sound from all directions, intelligently prioritise and provide the lead speaker's voice by suppressing unwanted noises, including speeches of other people. This device selects and alternates voices between speakers automatically using voice algorithms. Additionally, the user has the option of further fine-tuning the acoustic parameters as needed through a smartphone interface. This paper describes the development and functions of this new Smart Hearing Aid.