飛行機器人狀態估計使用被動無線感測器網路

Flying Robot State Estimation Utilizing Passive Wireless Sensor Networks

在開發送貨用的飛行機器人當中，機器人導航為其中重要的一個環節，而要實現導航的關鍵在於具有準確的機器人狀態估計。在不考慮使用GPS、超音波或雷射測距儀之動態定位問題上，本文提出一個動態定位的被動估計方法，即考量機器人利用被動無線感測器網路，其中組成包含多個地標（Landmarks）（地標位置為已知且本身需要輻射電磁波或紅外線），以及機器人本身需要裝置角度感測器接收來自於多地標的訊號，用以量測方位角與俯仰角，以及利用僅有角度的量測數據與多方來源資訊融合技術，即利用混合座標系濾波技術以及遞迴式估計融合演算法，進行機器人自主導航動態定位。本文利用兩座地標、預先規劃的機器人飛行路徑與機器人本身位置、速度的估計進行比對，運用蒙地卡羅電腦模擬技術，重複實驗，以驗證本方法之有效性。從模擬結果可知，本文所建議的演算法，適用於機器人導航問題上。 Precise navigation is one of the most important issues in the development of the shipping service of flying robots. The key to accomplish navigation relies on accuracy of positioning of the robots. The present research proposes a method to the robots in state estimation using passive wireless sensor networks under the situations where GPS, ultrasound or laser distance sensor are not to be considered as mobile positioning tools. The robots utilize the passive wireless sensor network including multiple landmarks and angular sensors installed on the robots. The on-board sensor receives angular signals (azimuth and elevation) from the landmarks, and then based on a group of hybrid coordinate (HC) filters a recursive estimation fusion algorithm is developed to estimate both position and velocity of the robots for autonomous navigation. Simulation results of proposed algorithm show that the convergence of estimation errors is greatly accelerated using only two landmarks. The algorithm markedly improves the estimation accuracy as well.

Precise navigation is one of the most important issues in the development of the shipping service of flying robots. The key to accomplish navigation relies on accuracy of positioning of the robots. The present research proposes a method to the robots in state estimation using passive wireless sensor networks under the situations where GPS, ultrasound or laser distance sensor are not to be considered as mobile positioning tools. The robots utilize the passive wireless sensor network including multiple landmarks and angular sensors installed on the robots. The on-board sensor receives angular signals (azimuth and elevation) from the landmarks, and then based on a group of hybrid coordinate (HC) filters a recursive estimation fusion algorithm is developed to estimate both position and velocity of the robots for autonomous navigation. Simulation results of proposed algorithm show that the convergence of estimation errors is greatly accelerated using only two landmarks. The algorithm markedly improves the estimation accuracy as well.