Abstract
The object of research is the computer system of unmanned aerial vehicles (UAVs). Most modern UAVs are based on 2- or 4-core single-chip processors, between which between which the OS automated scheduler tries to evenly distribute computational tasks . One of the most problematic places in the described process is that the first core can instantly become extremely congested in the event of an urgent task from the UAV control system. Therefore, the subject of research is complex indicators of the state of the processor cores for various algorithms of task distribution proposed between the cores of a multicore single-chip processor. In the course of the research, methods for simulating the dispatching of tasks processed by the UAV computer system based on a quad-core single-chip processor are used. The expediency of using the energy of the measuring signal from piezoelectric sensors for partial compensation of the consumed energy by the UAV computer system is investigated and justified. To evaluate the effectiveness of measures taken to increase the efficiency of the use of computer system components, the HWMonitor utility is used. According to the research results from the developed certain optimal algorithm, which differs from the others by reserving the resources of the 1st core of a multi-core single-chip computing processor for calculations of primary importance. The use of such calculation algorithm provides an increase in flight time by 3.1 minutes and increases the range of professional tasks by 1.3 min (for UAV DJI Phantom 4). In comparison with similar known solutions, the proposed algorithm improves the UAV stable behavior in critical applications (loss of ground control, the occurrence of obstacles, the impossibility of obtaining GPS coordinates in the areas of radio electronic warfare, etc.).
Highlights
The computing systems of most unmanned aerial vehicles (UAVs) are based on single-chip 4-core 32-bit processors ARM Cortex, Qualcomm Snapdragon, etc. [1, 2]
In the case under investigation, additional difficulties arise because the .NET stream that is run by the Common Language Runtime (CLR) does not correspond to the OS thread, and only the threads of the operating system can bind to the cores [19, 20]
The proposed approach in the complex significantly improves the efficiency of using UAVs based on a multi-core calculator and improves the conditions for its interaction with the ground command center
Summary
The computing systems of most unmanned aerial vehicles (UAVs) are based on single-chip 4-core 32-bit processors ARM Cortex, Qualcomm Snapdragon, etc. [1, 2]. The computing systems of most unmanned aerial vehicles (UAVs) are based on single-chip 4-core 32-bit processors ARM Cortex, Qualcomm Snapdragon, etc. The number of computational flows in them is equal to the number of physical cores, between which tasks can be distributed (orientation in space, traffic control, receiving and processing indicators from sensors, etc.). For a computer system with limited resources (computing or power supply), the following problem arises: the operation of an automatic scheduler of operating system (OS) flows leads to an unjustified processor load. For mobile computers on which UAV construction is based, it is often more important to economize in power consumption (in practice – battery life) than to attract all available cores in a single-chip computing processor (CPU) in the computational process. The actual task is to improve the efficiency of using UAVs by developing better, than automatic, algorithms for the operation of a multi-core processor, on the basis of which a UAV computer system is built
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
