This study explores the motivations behind integrating TinyML-based voice assistants into daily life, focusing on enhancing their user interface (UI) and functionality to improve user experience. This research discusses real-world applications like smart home automation, visually impaired assistive technologies, and healthcare monitoring. This review acknowledges various problems and helps us understand why TinyML exerts such significant implications in numerous domains. Researchers derive solutions from this study on how voice assistants integrated with TinyML can effectively analyze and adjust to user behaviour patterns in real-world scenarios, thereby enabling the delivery of dynamic and responsive content to enhance user engagement. The article also focused on limitations while implementing TinyML. Researchers will understand the detailed issues that are unavailable in most papers. This work explores features that can be embedded in voice assistants, like smart home automation, smart watches, smart glasses for visually impaired people, etc., using TinyML. A comparative review of current methods identifies areas of research gaps such as deployment difficulties, noise interference, and model efficiency on low-resource devices. From this study, researchers can directly identify the research gap with minimal effort, which may motivate them to focus more on solving the open problems due to optimize the problem identification time.
TinyML-based voice assistants enhance everyday life by improving user interfaces and real-world functionalities, While researchers focus on improving user adoption of voice assistants, fewer studies explore their integration with emerging technologies for expanded functionalities, such as smart home automation and healthcare monitoring. This voice assistant system helps consumers in various ways; in smart home assistants; in Smart Healthcare, etc. Researchers were also able to identify depression detection using multi-modal techniques using Speech and EEG signals. Artificial Intelligence (AI) is being applied in a wide range of research fields and has been shown to be an innovative tool for solving a wide range of research issues. Nevertheless, there is a cost associated with the massive processing needed to train AI systems. Driven by the need to lower the cost, carbon footprint, and energy use of the machines running ML techniques, TinyML is now seen as a viable alternative to artificial intelligence that focuses on applications and technology for incredibly low-profile devices. There are several exciting fields in which TinyML can have a significant influence. Anomaly detection plays a crucial role in industry by helping to minimize delays for repairs and boosting production efficiency. By implementing machine learning algorithms at the edge, it is feasible to continually monitor and evaluate the noise that the machine produces while it is in operation, which may indicate a potential malfunction. Real-time analysis of various parameters, such as noises or vibrations, can assist in saving time while replacing or fixing faulty equipment without causing further delays. The Web of Animals is one of the most recent study areas in the environment whereby sensors have been widely used. Most researchers still need help understanding animal behaviour. Studying animal behavior through continuous observation for short periods can be a challenging practice. The IoT, particularly TinyML, can significantly eliminate the need for this laborious work. Gaining more in-depth understanding of animal life and anticipating potential dangers might be beneficial. An elephant is fitted with a collar in the elephant's TinyML project uses GPS to track the elephant's movements in real-time. The implanted sensors gather pictures of its surroundings, which TinyML continually processes and analyzes to forecast occurrences surrounding each animal. While a motion sensor is utilized to assess the elephant's movement further, other machine-learning models may also be employed to comprehend and detect the elephant's mood. TinyML opens up new options and provides fresh perspectives on sustainable development. TinyML lowers latency so real-time applications, such as voice and picture recognition, may be implemented at the data source. Additionally, TinyML models may function without an internet connection-something that is not possible in a cloud environment. Because TinyML processes data without requiring it to leave the device, it dramatically enhances user privacy and conforms with data protection laws. The scope of the paper is to present more information about the TinyML feature with voice assistant, which is embedded in small devices with IoT.
Although TinyML provides considerable benefits, including low power usage, low latency, and increased privacy, it also comes with some trade-offs compared to cloud-based systems. One of the biggest drawbacks is the lower computational capability of microcontrollers, which limits the complexity of models that can be used. In contrast to cloud-based voice assistants that utilize large computational resources for deep learning, TinyML-based systems have to depend on optimized, light models that strike a balance between accuracy and efficiency. TinyML devices also come with limited memory and storage, which creates difficulties in dealing with large datasets and real-time adjustment. Yet these trade-offs are offset by advantages like enhanced data security-because processing is local instead of on remote servers-and reduced energy use, which makes TinyML ideal for edge applications in resource-scarce environments.
This study aims to evaluate the effectiveness of TinyML-based voice assistants by analyzing key performance metrics, including accuracy, computational efficiency, and power consumption, across different hardware platforms and deployment scenarios. Hypothesis TinyML-based voice assistants can achieve comparable accuracy to cloud-based systems while significantly reducing energy consumption and enhancing user privacy, making them viable alternatives for real-world applications.