150 likes | 171 Views
Compressed sensing in data collection. Yiying Zhao. Outline. Introduction Compressed sensing Network model Future work. 1. Introduction.
E N D
Compressed sensing in data collection Yiying Zhao
Outline • Introduction • Compressed sensing • Network model • Future work
1. Introduction • Energy efficiency of data collection is one of the dominating issues of wireless sensor networks(WSNs), especially when considering data collection. Because data collection will cause serious traffic load at the sink node. • The newly developed technique, compressed sensing, provides another method that allowing network recover the signal with high probability from far fewer samples than original dimension.
2. Compressed sensing • When reconstruct signal from v, we solve the following problem • There are many methods to solve this problem, such as basis pursuit algorithm and matching pursuit algorithm.
3. Network model • In this model, I deploy node intentionally, which means I already know the connection among these nodes. I also neglect the transmission loss. It extremely simplifies the whole model. • The model contains ones sink node to collect data and other sensor nodes to gather data.
3. Network model • Step 1 • Divide the nodes set (apart from the sink node) into two sets, A and B. the sum of those nodes are n. • A set contains nodes which receive data less than k-1. • B set contains nodes receiving data more than k.
3. Network model • Step 3 • Reconstruction
3. Network model • In reconstruction part, we are about to recover the data transmit from H1, H2, H3. • In the part, we choose the matching pursuit algorithm.
4. Future work • 1. do some simulation • 2. take more factors into consideration