Input: The unique ID of current node ID_c |
The node Type of current node Type_c |
Define: Localization Request Packet LRP |
Location Packet LP = {ID, Position, Type} |
Reference List RL = [IDi, pi, di] |
Localization Confidence Score LCS |
Estimated Position EP |
Confidence List CL = [ID, EP, LCS, Type] |
Localization Decision LD |
Procedures: |
1: send an LRP |
2: If received an LP: |
3: parse the LP to obtain ID, position, Type |
4: obtain the RSSI |
5: if Type==reference: |
6: obtain distance estimation d via Eq. (1) |
7: append ID, position, d, to RL |
8: If sizeof (RL) > =3: |
9: calculate current geometric features via Eqs. (13) (14) and (15) |
10: calculate the LCS of this localization by Eq. (19) |
11: IfLCS > LCS in the last CL |
12: input the features to one-class SVM classifier to |
obtain a decision result LD |
13: IfLD==localization: |
14: estimate the location of current node ID_c by |
Eq. (10). |
15: change the type of current node to pseudo-anchor |
16: append [ID_c, EP, LCS, pseudo-anchor] to CL |