WSN无线传感器网络复习

无线传感器网络复习

3.Node Architecture

3.1 模拟信号数字信号转换

A vibration sensor outputs an analog signal with a peak-to-peak voltage of 5V at a frequency of 100 Hz.

(a) What should be the minimum sampling frequency, so that no information is lost during the digitization process?
答：使用奈奎斯特定理，无损塑模出原始模拟信号需要频率为2f，也即100*2=200Hz

(b) Suppose a resolution of 0.025V is required to detect an interesting event. What should be the resolution of the ADC in terms of bits to convert the analog signal to a digital signal?
答：我们有公式
$$Q=\frac{E_{pp}}{2^m}$$
其中，$Q$是分辨率，$E_{pp}$是波峰到波谷之间的差，m是表示这个模拟量需要多少bit。
根据上述公式，我们计算
$$2^m=\frac{5-(-5)}{0.025}=400$$
因此我们至少需要9bit才能表示该模拟量。

3.5 微控制器广泛应用的原因

While they are not the most energy-efficient solutions, microcontrollers are the predominant processors in wireless sensor networks. Explain some of the reasons.

答：Microcontrollers offer greater programming flexibility compared with the other types of smallscale processors; hence, they are useful for many applications.
微控制器的优点：结构紧凑，体积小，功耗低和低成本，使其适合构建计算不太密集的独立应用。同时使用C语言为编程，使用户能够不必了解过多底层的具体细节。
微控制器相较其他小规模处理器的优点就是它提供了更加方便灵活的编程，使用高级语言编程。

3.6 无线传感器网络节点不使用冯诺依曼结构的原因

Explain the reason why using the Von Neumann architecture is not efficient for a wireles ssensor node.

答：In the Von Neumann architecture, there is a single memory space for data as well as program code and a single bus interfaces the memory unit with the processor. This means each data and instruction transfer requires a separate cycle, as a result of which computation is slow in lowscale processing subsystems.
冯诺依曼体系结构中，指令与数据存储在一起，而主存与处理器之间只有一个简单的总线用于传输数据，因此每次数据传输都花费一个独立的串行时钟计时。这样工作相当低效。
因此WSN采用Super-Harvard (SHARC) 超-哈佛结构，其中，IO流与数据存储器件之间使用DMA机制，不需要处理器控制实现数据直接存取，加快了处理速度。

3.7 为什么不用并行总线

Why are parallel busses not desirable in a wireless sensor node?

答：Parallel busses require large space. This is difficult to accommodate in small sensor nodes.
体积限制

3.12 FPGA与ASIC异同

Explain the basic similarities and differences between a FPGA and an ASIC.

答：

Difference between ASICs and FPGAs mainly depends on costs, tool availability, performance and design flexibility. Table 3.1 summarized some of their differences.

3.13 超-哈佛结构的特点

Explain some of the distinct features of the SuperHarvard
architecture.
(a) It provides an internal instruction cache to store frequently needed instructions.（指令cache存储常用指令）
(b) Program memory can be used to store data.
(c) Direct data streaming from an external hardware into the data memory through an I/O controller is possible.（使用DMA）

4.Operating System

4.6 调度方式比较

Compare the following scheduling mechanisms:
(a) FIFO scheduling
(b) sorted queue
(c) round-robin

答：
(a) FIFO scheduling
In a FIFO scheduling mechanism a task is processed based on the first-in-first-out principle.

(b) Sorted queue
In a sorted queue scheduling mechanism, tasks in a queue are first sorted according to a set criterion, for example, based on their length.

(c) Roundrobin
In roundrobin schedulingmechanism, timemultiplexing is used to divide time fairly among competing tasks. The execution time is divided into several slots and each task is executed in a slot. When the slot is due, the task is set on hold and the next task is executed. This way, all tasks progress together towards their completion.

4.7 中断与中断句柄

What are interrupts and interrupt handlers?

答：An interrupt is an asynchronous signal (event) and is generated by a hardware or software componentwhich requires immediate handling.When the processing subsystemreceives an interrupt event, it transfer control to an interrupt handler, also called an interrupt service routine (ISR). An interrupt handler is a callback subroutine which is called by the operating system when an interrupt event is received. Interrupt handlers should register for the interrupt types they are interested in.

4.8 为什么大多数WSN的操作系统都自定义内核

Why do most operating systems in wireless sensor networks define a kernel?

答：A monolithic kernel with a small finger print enables dynamicmodule update and dynamic
reprogramming.

4.12 解释配置组件与模块的区别

Explain the difference between configuration components and modules in TinyOS.

答：A configuration component describes how different modules are interconnected to build
an executable service or application, whereas a module is an implementation of an
interface.

配置组件描述可执行的服务与应用之间的关系，而模块是一个接口的具体实现。

4.15 基于事件与基于线程的操作系统之间的不同，并讨论并在WSN上下文中讨论二者的优点与缺点

Explain the difference between eventbased and threadbased operating systems. Discuss some of the advantages and disadvantages of the two approaches in the context of wireless
sensor networks.

答：
In eventbased programming：
interaction between processes is based on events and event handlers. Tasks are executed to completion, unless they are interrupted by events. This way concurrency is supported and execution is efficient. Since only one task is executed at a time, longduration tasks may block shortduration
tasks, but this problem can be overcome by using a sorted queue scheduling.

In multithreaded programming：
multiple threads run concurrently. Threads can be suspended ensuring nonblocking operation. However, thread management introduces resource overhead on the operating system.

基于事件的并发处理机制更加高效，正在运行的事件只能能被其他事件打断。虽然可能会出现饥饿现象，但可通过sorted queue调度算法避免。
线程会挂起以避免阻塞，但线程管理会带来额外开销。

4.18 TinyOS中的基本概念

Explain the following concepts in TinyOS:
(a) commands
(b) tasks
(c) events

答：
(a) Commands：
Commands are nonblocking
requests for service.
(b) Tasks：
Tasks are monolithic processes that should be executed to completion
(c) Events：
An event is an occurrence of interest outside of a process and prompts the process to act (or handle the event).

6.Medium Access Control

6.2 基于冲突和冲突避免的MAC子层协议优劣

What are the advantages and disadvantages of contentionfree and contentionbased medium access strategies?
Can you think of scenarios where one would be preferable over the other?

答：
Contentionfree
medium access strategies avoid collisions by ensuring that when one device transmits, all other devices that could interferewith the reception of the transmitted data remain silent. Avoiding collisions can have positive impacts on communication latency and throughput. Most contentionfree approaches are based on schedules, indicating exactly when a node can transmit or must listen. This also facilitate power management strategies, e.g. using such a schedule, a device knows exactly when it can power down its radio.
On the other hand, contentionbased mediumaccess
strategies do not avoid collisions, but
instead provide mechanisms to recover from collisions. An advantage of such a strategy is that a device does not have to wait (and delay), but instead can transmit immediately. In networks with lowtraffic loads, this approachmay result in lower latencies.A contentionbased strategy may also be easier to implement, since no schedules are required. Besides the potential collisions (which may lead to increased latency and reduced throughput), it may bemore difficult to implement duty cycle techniques in a network using a contentionbased MAC protocol.

Contentionbased
access strategies may be preferable in networks with light traffic loads and unpre