s********k 发帖数: 6180 | 1 接收端收到两个来自不同MS的信号(s1,s2),在接收端的power分别为(p_s1, p_s2)如
果p_s1/p_s2>=capture_thresh
1. s1先到达,s2后到达,接收端将正确接收s1,drop掉s2.这个没有太大疑问吧
2.s2 先到达,s1后到。这样的话,接收端会不会drop掉s2,转而接受s1?
3. 条件同2, 但是carrier sensing thresh
sense但是由于RSSI不够,没法被正确接收(CRC error)。这样的话能否正确接收到后
来的s1信号? |
z*****n 发帖数: 7639 | 2 呵呵,看来搞MAC的还是不多啊。。。
【在 s********k 的大作中提到】 : 接收端收到两个来自不同MS的信号(s1,s2),在接收端的power分别为(p_s1, p_s2)如 : 果p_s1/p_s2>=capture_thresh : 1. s1先到达,s2后到达,接收端将正确接收s1,drop掉s2.这个没有太大疑问吧 : 2.s2 先到达,s1后到。这样的话,接收端会不会drop掉s2,转而接受s1? : 3. 条件同2, 但是carrier sensing thresh: sense但是由于RSSI不够,没法被正确接收(CRC error)。这样的话能否正确接收到后 : 来的s1信号?
|
z*****n 发帖数: 7639 | 3 1. s2 is regarded as noise in this situation
2. Theoretically, this can be done. But I doubt that in reality
the receiver circuit cannot be turned to s1 because:
each 802.11 frame contains a set of preamble bits which is
used by the receiver to synchronized its clock for correct
reception. Once the clock is synchronized, it will continue
the receiving phase until the channel becomes clear. So,
if s1 comes during the reception of s2, it will be regarded
as an overshouting noise
【在 s********k 的大作中提到】 : 接收端收到两个来自不同MS的信号(s1,s2),在接收端的power分别为(p_s1, p_s2)如 : 果p_s1/p_s2>=capture_thresh : 1. s1先到达,s2后到达,接收端将正确接收s1,drop掉s2.这个没有太大疑问吧 : 2.s2 先到达,s1后到。这样的话,接收端会不会drop掉s2,转而接受s1? : 3. 条件同2, 但是carrier sensing thresh: sense但是由于RSSI不够,没法被正确接收(CRC error)。这样的话能否正确接收到后 : 来的s1信号?
|
c****n 发帖数: 21367 | 4 大侠展开说说?:)
)如
到后
【在 z*****n 的大作中提到】 : 呵呵,看来搞MAC的还是不多啊。。。
|
z*****n 发帖数: 7639 | 5 大侠不敢当,混饭的玩艺而已。
【在 c****n 的大作中提到】 : 大侠展开说说?:) : : )如 : 到后
|
c****n 发帖数: 21367 | 6 CISCO gave us a topic which is quite similar to the OP's problem
"how to distinguish two almost simultaneously arriving signals,
(in PHY layer)"
p1/p2 may be used in this context or similar.
i agree with you, on MAC layer, beacuse of the pilot bits,
it is hard to replace s2 over s1 if s2 arrived later.
practically, if s2 is stronger than s1, i guess s1 will have
many errors instead of s2 started being received.
)如
到后
【在 z*****n 的大作中提到】 : 1. s2 is regarded as noise in this situation : 2. Theoretically, this can be done. But I doubt that in reality : the receiver circuit cannot be turned to s1 because: : each 802.11 frame contains a set of preamble bits which is : used by the receiver to synchronized its clock for correct : reception. Once the clock is synchronized, it will continue : the receiving phase until the channel becomes clear. So, : if s1 comes during the reception of s2, it will be regarded : as an overshouting noise
|
s********k 发帖数: 6180 | 7 这个问题后来查了查,来源是一些capture effect的文章。即便s1还在preamble接收期
间,s2到达,这个叫做preamble capture,大概是4db的extra就可以。如果s1已经在正
式frame body接收期间。s2到达,需要10db的extra才能正确接收。不过以上两种情况
都需要网卡支持,开启一种message in message的mode。
【在 z*****n 的大作中提到】 : 1. s2 is regarded as noise in this situation : 2. Theoretically, this can be done. But I doubt that in reality : the receiver circuit cannot be turned to s1 because: : each 802.11 frame contains a set of preamble bits which is : used by the receiver to synchronized its clock for correct : reception. Once the clock is synchronized, it will continue : the receiving phase until the channel becomes clear. So, : if s1 comes during the reception of s2, it will be regarded : as an overshouting noise
|
c****n 发帖数: 21367 | 8 问题在于,什么样的应用下s2比s1应该优先呢?密集布网?
【在 s********k 的大作中提到】 : 这个问题后来查了查,来源是一些capture effect的文章。即便s1还在preamble接收期 : 间,s2到达,这个叫做preamble capture,大概是4db的extra就可以。如果s1已经在正 : 式frame body接收期间。s2到达,需要10db的extra才能正确接收。不过以上两种情况 : 都需要网卡支持,开启一种message in message的mode。
|
z*****n 发帖数: 7639 | 9 this is not a problem of priority, it is about the efficiency:
Suppose s2>s1 and it arrives later...
if capture is disabled, s2 will collide s1 and nothing is received;
if capture is enabled, s2 will be successfully received.
【在 c****n 的大作中提到】 : 问题在于,什么样的应用下s2比s1应该优先呢?密集布网?
|
s********k 发帖数: 6180 | 10 关于情况3,有个问题还不是很清楚。一个节点只要sense到RSSI高于Carrier sense的
thresh,是不是就会开启接受模式开始接受?如果这个信号的RSSI高于RX thresh,但是
目的地址并不是当前节点,当前节点什么时候丢弃这个包?是PHY在接受完了preamble
之后就立即丢弃,还是PHY完整接受完上传给MAC之后,MAC来判断?我印象中是后者,
不知道对不对。同样CS
个包没法正确接收吧?
【在 z*****n 的大作中提到】 : 1. s2 is regarded as noise in this situation : 2. Theoretically, this can be done. But I doubt that in reality : the receiver circuit cannot be turned to s1 because: : each 802.11 frame contains a set of preamble bits which is : used by the receiver to synchronized its clock for correct : reception. Once the clock is synchronized, it will continue : the receiving phase until the channel becomes clear. So, : if s1 comes during the reception of s2, it will be regarded : as an overshouting noise
|
|
|
s********k 发帖数: 6180 | 11 关于这个,我在一个其实简单的方针里面都能做到这点。就4个节点。
【在 c****n 的大作中提到】 : 问题在于,什么样的应用下s2比s1应该优先呢?密集布网?
|
c****n 发帖数: 21367 | 12 oh, that's the point. :)
【在 z*****n 的大作中提到】 : this is not a problem of priority, it is about the efficiency: : Suppose s2>s1 and it arrives later... : if capture is disabled, s2 will collide s1 and nothing is received; : if capture is enabled, s2 will be successfully received.
|
s********k 发帖数: 6180 | 13 right,所以capture effect能够提高throughput,不过也有人质疑其能量消耗以及
near far的问题。就能量问题,从电路角度肯定比不开capture effect消耗大,不过从
系统角度未必。因为省掉了很多不必要的重传。第二个问题我觉得CCA的存在也会解决
绝大部分near far的问题
【在 z*****n 的大作中提到】 : this is not a problem of priority, it is about the efficiency: : Suppose s2>s1 and it arrives later... : if capture is disabled, s2 will collide s1 and nothing is received; : if capture is enabled, s2 will be successfully received.
|
c****n 发帖数: 21367 | 14 CCA = ?
【在 s********k 的大作中提到】 : right,所以capture effect能够提高throughput,不过也有人质疑其能量消耗以及 : near far的问题。就能量问题,从电路角度肯定比不开capture effect消耗大,不过从 : 系统角度未必。因为省掉了很多不必要的重传。第二个问题我觉得CCA的存在也会解决 : 绝大部分near far的问题
|
s********k 发帖数: 6180 | 15 clear channel access。11,15.4 发送前必要的动作
【在 c****n 的大作中提到】 : CCA = ?
|
z*****n 发帖数: 7639 | 16 I think so, address is called MAC address, it should be
checked at MAC layer.
preamble
【在 s********k 的大作中提到】 : 关于情况3,有个问题还不是很清楚。一个节点只要sense到RSSI高于Carrier sense的 : thresh,是不是就会开启接受模式开始接受?如果这个信号的RSSI高于RX thresh,但是 : 目的地址并不是当前节点,当前节点什么时候丢弃这个包?是PHY在接受完了preamble : 之后就立即丢弃,还是PHY完整接受完上传给MAC之后,MAC来判断?我印象中是后者, : 不知道对不对。同样CS: 个包没法正确接收吧?
|
z*****n 发帖数: 7639 | 17 clear channel assessment :-)
【在 s********k 的大作中提到】 : clear channel access。11,15.4 发送前必要的动作
|
c****n 发帖数: 21367 | 18 另一个问题在于,s2强就一定是需要收的信号么?
会不会导致fairness的问题?
【在 s********k 的大作中提到】 : right,所以capture effect能够提高throughput,不过也有人质疑其能量消耗以及 : near far的问题。就能量问题,从电路角度肯定比不开capture effect消耗大,不过从 : 系统角度未必。因为省掉了很多不必要的重传。第二个问题我觉得CCA的存在也会解决 : 绝大部分near far的问题
|
s********k 发帖数: 6180 | 19 不一定,这个实际就是我现在面临或者想解决的问题。至于是不是有fairness问题。
肯定会有,而且由于capture effect对于到达顺序比较敏感。所以中间还有很多问题。
比如接受信号先到,干扰后到,只需要接收信号比干扰信号强就可以了。但是干扰信号
先到,接收信号后到,接受信号需要比干扰信号强Ndb才行。所以即使设置不变,都会
有fairness问题
【在 c****n 的大作中提到】 : 另一个问题在于,s2强就一定是需要收的信号么? : 会不会导致fairness的问题?
|
s********k 发帖数: 6180 | 20 sorry, you are right:)
【在 z*****n 的大作中提到】 : clear channel assessment :-)
|