APPLICATION OF MACHINE LEARNING METHODS TO LOCALIZE QUADCOPTER SENSOR FAILURES
Annotation
The problem of localizing failures of sensors (accelerometer and gyroscope) of an unmanned aerial vehicle of the "quadcopter" type is considered. An algorithm is developed that provides the ability to detect and classify quadcopter sensor failures using machine learning methods. To solve the problem, the following machine learning methods were used: logistic regression, random forest method, LASSO and ridge regression, as well as elastic net. Experimental results obtained in the course of computer simulation confirm the efficiency of the proposed algorithm. A comparative analysis of the used methods of machine learning is performed.
Keywords
Постоянный URL
Articles in current issue
- SPECIAL PROPERTIES OF HAMMING CODES THAT APPEAR WHEN SYNTHESIZING SELF-CHECKING DIGITAL DEVICES
- CHOOSING THE OPTIMAL CONFIGURATION OF THE ONBOARD COMPLEX OF THE SPACECRAFT TO RESTORE ITS OPERABILITY
- FEATURES OF THE IMPLEMENTATION OF A DISTRIBUTED VIRTUAL MACHINE WHEN BUILDING A GENERAL-PURPOSE COMMUNICATION INFRASTRUCTURE
- DEVELOPMENT AND IMPLEMENTATION OF ELEMENTS OF AN INFORMATION AND ANALYTICAL PLATFORM FOR SOLVING TRANSPORT AND LOGISTICS PROBLEMS
- APPLICATION OF MACHINE LEARNING METHODS TO LOCALIZE QUADCOPTER SENSOR FAILURES
- MONITORING THE DEGREE OF WOOD THERMAL DAMAGE BY MEASURING ITS COLOR CHARACTERISTICS
- EXPERIMENTAL STUDY OF ACOUSTIC EMISSION PARAMETERS DURING CYCLIC TESTING OF METAL PRODUCTS OF ADDITIVE MANUFACTURING
- USE OF A FIBER OPTIC SPECTROMETER TO ASSESS THE MINIMUM NUMBER OF MALIGNANT CELLS BY DIFFUSE REFLECTANCE SPECTRA IN THE NEAR INFRARED REGION
- METHODOLOGY FOR CREATING PARAMETRIC MODELS OF PARTS FOR AUTOMATED DESIGN OF TECHNOLOGICAL PROCESSES
- ANOMALOUS SPECTRUM COMPRESSION AS A MEANS OF OBTAINING RADIATION OF SUPER-PLANCK INTENSITY