# 資料庫Proposed sports activity monitoring using big data visualization (SAM-BDV) method

> 「computer-assisted mobile sensing networks」 were suggested in athlete health tracking systems utilizing IoT.

:bookmark:傳輸方式：
* <Font color="FF0000">Body Sensing Node (BSN) </Font>collects biodata from sports for analogue front sensor nodes and other detectors.
* Through standard Bluetooth **low energy connection and Wi-Fi protocols**, the Body Sensor transmits detected information to an entrance.(**The gateways** then send the data to a distant central server from body sensing nodes.)
* Furthermore, the server presents **visualization and diagnostic imagery** on linguistic(語言) or visual dashboards(視覺) for medical practitioners(醫療人員). The proposed model **collects health information** through the Cloud Infrastructure, BSN, Cloud Gateways (CG)
### Figure 1. shows the architecture of the proposed SAM-BDV model.
* An **analysis model and utilization** of the portable sensors show how the expense of resource calculation **can be lowered** while preserving(維持) health needs for medical data in the fog and cloud distributing environments.
* The assessment(評估) system is based on a queue to estimate the **minimal computer resources** required for a service level assurance (SLA).
* <Font color = "FF0000">Weakness:
(1)examining wide-ranging data techniques and employing data from previous aggregates(過去及現有資料的整合)
(2)the connection between handheld(手持)(mobile) and centralized(集中式) sensors might result in
certain delays and blockages(阻塞)
</Font>
* **Confidential material can be transferred to Cloud Environment(Private Cloud) from GW**, which does not interface with Edge devices or cloud hosting for public cloud interaction.(私密性)
* Fog-assisted Distant Management System shows significant energy usage **reductions and minimal accuracy mistakes**.and conforms (遵守)to the IoT devices in bandwidth(頻寬) and power usage that achieve high accuracy.
* **It isn’t easy to utilize a centralized technique, which analyses the Big Data techniques more widely and uses past aggregate information.**
* This approach seeks to create biological sensors and computer **capability** to allow patients of sophisticated applications. This study involves implementing novel overall healthcare programs of the computing compatibility of smart devices utilizing IoT and fog-supported platforms.

### * Case 1: Queuing Model(排隊理論模型)
1. 用arrival rate(到達率) λi來表示高斯分布，the probability distribution is geometrically dispersed(幾何離散)，藉由1/λi算出所有健康照護資訊服務的平均inbound time(呼叫總時間/呼叫總次數)(幾秒會有一個到達)，相加可獲得response time.
2. The arrival rate is denoted as λi(平均到達的顧客數 個/秒). The healthcare data received in each model part is processed using a first-come, first-serve technique (FCFS).
3. The cloud platform and gateways (CW and H) are modelled as an **endless waiting buffer in M/M/1**(隊列長度無限制) to avoid losing health information between the web of the objects processor model and between public cloud storage systems and fog layers.
### Figure 2. Fog-Cloud environmental architecture of the customer condition monitoring.

* The processing times of H and CW are dispersed as probability distributions with an average rate of 1/μH and 1/μC accordingly similar and separately(每客平均服務時間 T= 1/μ).
* Let us assume an N/N/1/C queuing with identical service time μF per each cloud environment (有系統容量限制)and an N/N/D endless queuing buffer to assess cloud power without limits(沒有系統容量限制)
* The N/N/d/l queue illustrates any individual node**(multiple)** in a **personal cloud infrastructure**. The random distribution variable with rate 1/ μ1 is same and unbiased in each node operation time. (平均服務時間相同)
* **But the healthcare loading data entered throughout the period**, which have been changed, depends **on the number of major services** necessary to generate dynamism and remove computational assets (clouds and fog computing).(服務台數量增加，但可能會影響到loading data)
* mission readiness testing (MRT)(任務準備度測試) .The MRT has assessed during 2000–20,000 s healthcare information queries.For additional services with delivery options, the Fog computer nodes route IoT information and community healthcare data is expedited faster. (IOT更快)
* The **N/N/1/C queuing** output simulates fog computing using the Poisson(泊松分布) technique with the Barker's theory activation functions. (用的是最經典的MM1排隊模型)(排隊服務時間呈指數分布一個服務台有限的服務容量)
* The Portal and Clouds Gateway (H and C) have an unlimited awaiting buffer queue of an N/N/2 queue to cover all incoming data working load and near-endless cloud capabilities. The **likelihood(probability)** requirements for health information in H is expressed in Eq. (2)
ρ:平均利用率，一段相當長的時間內可測得 = λ/μ≤1
pix =>(在稳态条件下的概率模型穩態機率為何)
(穩態過程處於狀態 i（i=>包含 i 個服務中客戶）並達到穩態概率為)
(應用於伺服器共用)
The likelihood condition and the load are expressed πx and σ for the xth node.The H queue calculates the **average number of healthcare data queries pending and expressed** in Eq. (4)
(在佇列中等候服務的人數)
* 執行完上述公式後discard 服務，Fog nodes transport the data to the IoT devices.
* 追求穩態，提高可能性
* **The servicing time for the Clouds Gateway** is the same and independently exponential dispersed parameter with a mean 1/μC rate(相同). The likelihood of a steady-state can be found that cognitive healthcare data requirements in C are expressed in Eq. (8),
* ### Case 2: Theoretical model on Fog Computation

備註：
* SLA(服務等級保證):Latency(延遲)、Packet Loss(封包丟失)、Jitter(抖動)(接收封包的延遲變化)、Throughput(傳輸量)
* FCFS (First-Come First-Served)FCFS 先到先處理排程 是所有排程演算法中最簡單的，所有的任務都會被放進 Queue 當中，最先到的任務會優先被處理，直到處理完畢或是該任務主動讓出使用權，作業系統才會處理下一個任務。
* MM1特性(四步驟 到達->排隊->服務機構->離去)
到達時間卜瓦松過程（Poisson process）
服務時間是指數分佈（exponentially distributed）
只有一个服务台（server），遵循先到先服务规则
隊列長度無限制
可加入隊列的人數為無限
a/b/x/y/
a=>顾客到达间隔时间的分布
b=>服务时间的 分布
x=>服务台数目(例如服務台員工數)
y=>系统容量限制
M(到达间隔时间為指数分布)
D(確定的間隔時間)
https://blog.csdn.net/qq_29831163/article/details/89735349
λ：平均到達的顧客數（單位時間平均到達率，個/秒）
μ：平均服務的顧客數（服務率、離開率,個/秒），每客平均服務時間 T= 1/μ(可以表示節點node的服務能力),