大規模非同步多用戶之通道使用的延遲和活動偵測

Abstract

在大規模物聯網（IoT）中，為了省電，許多設備處於不傳送數據的狀態，每段時間中僅有少數用戶處於活動狀態。為了保證可靠的數據傳輸，檢測這些物聯網設備中的活動用戶為哪些設備及其路徑延遲非常重要。為了應對這種情況，我們利用熱帶演算法來幫助我們以非常少的正交資源達成目標，這與系統中的總用戶數相比是非常少的。此外，為了減少延遲，我們採用長度有限的碼，可以使得接收端在不必等待所有碼到達的情況下知道活動用戶及其路徑延遲。我們還發現了接收碼的有限時間期限，以及其與碼和路徑延遲之間的數學關係。最後，即使物聯網系統中的總用戶數大幅增加，我們仍能輕鬆地推導出我們所要用到的碼字，同時保持數據傳輸的可靠性。根據我們在無噪聲通道下的模擬結果，隨著正交資源數量的二次方增長，我們的設計在每個總用戶數下均無檢測錯誤。

In large IOT devices, in order to reduce power, most devices are silent and only few users are active for each time frame. It is important to detect who are the active users in our IOT system and their path delays to guarantee reliable data transmission. We leverage tropical-arithmetic to help us reach our goal with very less orthogonal resources comparing to the number of total users in our system. Moreover, to reduce latency, the codeword we adopt is finite length and the receiver obtains message without having to wait for all codewords to arrive. We also found the relationship between finite receiving deadline and codewords’ length and path delay. Lastly, we are able to derive our codebook easily when the number of total users increase dramatically in our IOT system, while still remaining reliable data transmission. According to our simulation results for noiseless channel, our design has zero detection errors for every number of total users, as the number of total users increases quadratically according to the number of orthogonal resources.