Task scheduling

Abstract

This article presents a new approach to the design of task scheduling algorithms, where system-theoretical methodologies are used throughout. The proposal implies a significant perspective shift with respect to mainstream design practices, but yields large payoffs in terms of simplicity, flexibility, solution uniformity for different problems, and possibility to formally assess the results also in the presence of unpredictable run-time situations. A complete implementation example is illustrated, together with various comparative tests, and a methodological treatise of the matter.

Tiny Neural Networks for Environmental Predictions: An Integrated Approach with Miosix

“…We used a single neural network with 3 FP32 inputs, one for each average pressure value associated with a timeband; • MCU: for our project we chose a NUCLEO-F401RE based on the MCU STM32F401RET6, which is an ARM cortex M4 with 512 KB of FLASH memory and 96KB of SRAM. Through a multi threading mechanism, enabled by Miosix [14], and the use of interrupts, we have ensured that the DTNN and the sensor data acquisition are synchronized and power-efficient during the execution; • Bridge component: since the MCU chosen for our project does not have natively any integrated data display device, we sent the measured data and predictions made by the DTNN to a Rapsberry Pi via USB, which in turn, using a Python script, forwards them to the data visualizer. This approach was used in order to log the gathered data and validate the performance of the prototype; in a real deployment the data could be consumed locally in the microcontroller; • Data visualizer: in order to graphically display and perform some analysis on the performance of our system we used an open source time series database: InfluxDB 2.0 Cloud.…”