Edge LSTM Towards Deep and Sequential Edge Computing for IoT Applications
ABSTARCT :
The time series data generated by massive sensors in Internet of Things (IoT) is extremely dynamic, heterogeneous, large scale and time-dependent. It poses great challenges (e.g. accuracy, reliability, stability) on the real-time analysis and decision making for different IoT applications. In this paper, we design, implement and evaluate EdgeLSTM, a unified data-driven system to enhance IoT computing at the network edge. The EdgeLSTM leverages the grid long short-term memory (Grid LSTM) to provide an agile solution for both deep and sequential computation, therefore can address important features such as large-scale, variety, time dependency and real time in IoT data. Our system exploits the advantages of Grid LSTM network and extends it with a multiclass support vector machine by rigorous regularization and optimization approaches, which not only has strong prediction capability of time series data, but also achieves fine-grained multiple classification through the predictive error. We deploy the EdgeLSTM into four IoT applications, including data prediction, anomaly detection, network maintenance and mobility management by extensive experiments. Our evaluation results of real-world time series data with different short-term and long-term time dependency from these typical IoT applications show that our EdgeLSTM system can guarantee robust performance in IoT computing.
EXISTING SYSTEM :
? Existing edge nodes have limited processing capability, we also design a novel offloading strategy to optimize the performance of IoT deep learning applications with edge computing.
? Most existing deep learning applications (e.g., speech recognition) still need to be cloud-assisted.
? Edge computing is proposed to move computing ability from centralized cloud servers to edge nodes near the user end. Edge computing provides two major improvements to the existing cloud computing.
? In the performance evaluation, we test the performance of executing deep learning tasks in FEC architecture for edge computing environment.
DISADVANTAGE :
? In this paper, we propose leveraging the IoT virtualization concept and apply it to AI-powered IoT devices in order to tackle the aforementioned issues.
? Unlike in centralized deployments, distributed AI solutions may suffer from interoperability issues, due to fragmented and mainly application-specific solutions.
? To circumvent this issue, it is crucial to set up mechanisms to identify and discover AI components and build intelligent applications upon them, while efficiently using network and computing resources.
? The update can be issued for instance by monitoring the accuracy levels achieved in performed inference procedures or upon feedback received by the consumer applications.
PROPOSED SYSTEM :
• Edge computing is proposed to move computing ability from centralized cloud servers to edge nodes near the user end.
• Proposed model uses the flexible framework to provide optimized assignment of the task to the edge layer and cloud layer which reduces the rigidity.
• Proposed model combines the deep learning and flexible architecture and produce better performance using multiple agents.
• Proposed model performs deep learning in FEC architecture using multiple agent for edge computing.
• These data include structured data, such as temperature, vibration or multimedia information, such as video, images, and sounds.
ADVANTAGE :
? Relying on the local edge capabilities, instead of moving data to powerful remote data centers, can boost the system performance by reducing the data traffic load traversing the core network and thus ensuring low-latency access to context-aware cloud-like services.
? We report results measuring the performance achieved when running an object detection inference task.
? Mobile Net is designed for efficient inference in various mobile and embedded vision applications.
? This approach makes the discovery of IoT services easier, since metadata are used to index the virtual objects.
? It has been used in commercial implementations and within the FIWARE initiative.
|