基因演算法在柱形物件搬運機器人節能軌跡規劃之應用研究

Abstract

近年來隨著科技的進步，機器人及機械手臂的應用越來越廣，加上人力資源成本的上漲，機械手臂已經變成取代勞力的最佳方法；而近年來節能減碳的意識抬頭，省能的機械裝置已成為最近熱門的議題，柱狀物的夾持與搬移的規畫研究需考慮夾持點與移動(含迴轉)等動態參數的影響，在機器人實際應用與控制方面是非常具體而實用的課題。本研究選擇柱狀物為夾持件進行機器人軌跡規劃與節能效率關聯的探討。為了讓柱形物件搬運型機械手臂能達到降低耗能的效果，本論文分兩個部分達成，第一部分提出藉由重力平衡的機構使機械手臂系統在運作過程中可以隨時調整到重力平衡狀態降低重力對耗能的影響，而朝向手臂在靜止狀態時制動器的扭力輸出可以降到零扭矩的理想目標探討。第二部分則是使用最佳化數值方法 - 基因演算法規劃最節能的路徑與控制軌跡；基因演算法就是模擬遺傳學的基因演變法則進行演算；類似達爾文的物種進化論敘述的物競天擇、強者生存，而弱者會被自然淘汰的生存競爭法則。基因演算法模仿自然界生物進化機制發展起來的隨機全域搜索和優化方法，其本質是一種高效、並行、全局搜索的方法，他能在搜索過程中自動獲取和累積有關搜尋空間的資訊，並自適應地控制搜索過程以求最佳解。 本研究測試範例是以兩軸的機械手臂，抓舉一柱狀物體由水平的平台上翻轉傳送至垂直狀態的另一放置處；整個分析模擬過程中以省能為主要目標進行各種抓持點與運動軌跡等設計的測試。比較重點在於求出最低耗能的移動軌跡，使用基因演算法求出在搬運柱形物體的過程中，整個系統的最低耗能之移動軌跡，亦即控制軌跡最佳化的規劃；最後將整個節能軌跡設計與規劃系統實際應用在本實驗室(NCTU CIDM Lab.)開發的三軸五桿機械手臂上，並測量相關數據做為驗證，本研究成果將可用來求相似機械手臂移動時的最低耗能軌跡，有助於大量節省能源的消耗提高工作效能。

With the advance of science and technology, the use of robots and mechanical arms has become widespread. Also, mechanical arms have become the best way to replace labour because the increasing on human resource costs. In addition, energy-saving machinery has become one of the hottest issues due to the awareness of CO2 emission reduction and energy saving in recent years. The clamping and moving of pillars need to consider the impact of the dynamic parameter of clamping point and mobile, including rotation. It has a very practical and specific in the practical application. The main purpose of the paper is discussing the comparison of energy dissipation to different routes that mechanical arms consume when it clamps pillar. In order to reach the effect of reducing energy consumption, the paper divides into two parts. The first part presents the design ideas of mechanical arm system to adjust to the condition of gravitational equilibrium at any time during the operation, according to the system of gravity balance. While the mechanical arm is in the condition of rest, the torque output of the servo can be decreased to zero, which is the ideal target. The second part uses numerical methods – Genetic Algorithm to plan and control the best way of saving energy. Genetic Algorithm is to calculate by simulating the rules of genetic inheritance. Charles Robert Darwin’s theory – On the Origin of Species, explains populations will change to adapt their environments and after a period of time, they will form to another new species. Genetic Algorithm is a random search and optimization method from simulating all species in nature that developed from evolution. It is a high efficiently way. Also, it can gain and accumulate information which relates searching spaces automatically and find the best solution from searching. The target of the two-axis mechanical arm that we design is to take and send a pillar from horizontal platform to another vertical place. Energy saving is the main goal that we proceed and test every design in the process of the whole analysis. Another main point of the plan is to calculate the moving path which costs the lowest energy. In this part, we are going to use Genetic Algorithm to find the moving track which expends the minimum energy in the entire system from the process of carrying the pillar. That is, the optimizing plans of controlling the path. The system will be practical use on the three-axis mechanical arm and measure the related data to be verified in NCTU CIDM Lab at last. The result of the research can be used to look for the lowest energy dissipation path from similar mechanical arms.