Precision landing using an adaptive fuzzy multi-sensor data fusion architecture

Abstract

The positional inaccuracies associated with the GPS/INS measurements make the terminal phase of the normal GPS/INS landing system imprecise. To solve this problem, an adaptive fuzzy data fusion algorithm is developed to obtain more accurate state estimates while the vehicle approaches the landing surface. This algorithm takes the translational displacements in x and y from the mounted Optical Flow (OF) sensor and fuses them with the INS attitude measurements and the altimeter measurements. This low cost adaptive algorithm can be used for vertical landings in areas where GPS outages might happen or in GPS denied areas. The adaptation is governed by imposing appropriate assumptions under which the filter measurement noise matrix R is predicted. The R matrix is continuously adjusted through a fuzzy inference system (FIS) based on the Kalman innovative sequence and the applied covariance-matching technique. This adaptive fuzzy Kalman fusion algorithm (AFKF) is used to estimate the vehicle's states while landing is being commanded. AFKF results are compared with these obtained using a classical Kalman estimation technique. The AFKF algorithm shows better states estimates than its conventional counterpart does. Compared to prior landing systems, the proposed low cost AFKF has achieved a precision landing with less than 10 cm maximum estimated position error. Real precision landing flights were conducted to demonstrate the validity of the proposed intelligent estimation method.

Deep-Learning-Based Neural Network Training for State Estimation Enhancement: Application to Attitude Estimation

“…The lineraization process was performed at hover point where the attitude states are around 0°. The linearzation of the quadrotor dynamics is thoroughly illustrated in [22]. The process and measurement noise covarince matrices of the filter are described in equations (10)(11).…”

“…A fuzzy logic is integrated with the conventional Kalman to reduce the measurement noise and enhance the position estimation of capacitive touch panels [21]. Similarly, an adaptive multi-sensor fusion technique has been implemented to perform a precision landing for the RUAV system [22]. The fuzzy rules have been implemented to adapt the variances of the measurement covariance matrix R of the Kalman filter.…”

Hybrid Adaptive Computational Intelligence-based Multisensor Data Fusion applied to real-time UAV autonomous navigation

“…Focusing on the applications, Fuzzy logic is extremely flexible for the most diverse uses since it uses "membership functions" (MFs) that model and quantify the meaning of the symbols [50,70] for their deductive apparatus in an intuitive and close to the natural language way [54], through well-defined "ifthen" relationship rules [70]. A good example of the application of Fuzzy logic is its use to improve the accuracy of the landing procedure of an UAV [38]. Another is the joint use with Computational Vision tools for estimating the position of an UAV in real-time flight, based on landmark recognition, as seen in [52].…”

Distributed Error Correction of EKF Algorithm in Multi-Sensor Fusion Localization Model

“…Muniandi and Deenadayalan [14] used wheeled sensors, radar and GNSS as data acquisition sensors, and constructed a nonlinear real-time localization model by probability weighting method. Al-Sharman et al [15] used Kalman innovation sequence and covariance matching technology to continuously adjust through fuzzy inference system, and proposed real-time localization based on adaptive fuzzy Kalman fusion algorithm (AFKF). Plangi et al [16] proposed a real-time localization algorithm based on Kalman filter algorithm to solve the routing problem in real-time localization.…”