System and method for mobile location by dynamic clustering
A system and method for determining the location of a mobile device in a communications network. A set of network measurement reports (NMR) may be provided for a region in the network. NMRs in the first set may be clustered as a function of observed measurement information in the NMRs. A candidate NMR for a mobile device may be received, and the candidate NMR may then be associated with a selected cluster of NMRs. The candidate NMR and a second set of NMRs may then be clustered, and a location of the mobile device determined as a function of the clustered candidate NMR and second set of NMRs, where the NMRs in the second set are associated with the selected cluster.
1. A method of determining the geolocation of a mobile device operating in a region served by a communications network, the method comprising the steps of:
(a) determining, in observation space, an observation cluster of measurement data from a first set of network measurement reports (“NMRs”) from the communication network;
(b) comparing measurement data from a candidate NMR for a mobile device with the determined observation cluster of measurement data; and
(c) determining a location of the mobile device as a function of the compared measurement information,
wherein the NMR represents a single vector of observed measurement information, multiple vectors of observed measurement information, or a single vector representing multiple vectors of observed measurement information.
2. The method of claim 1 wherein the location is determined as a function of a location corresponding to the compared measurement information.
3. The method of claim 1 wherein the location is determined by:
(a) determining a centroid in the observation space based on the measurement data from the set of NMRs from the communication network and based on the measurement data from the NMR for the mobile device;
(b) calculating, based on the determined centroid, a weight for each of select ones of the NMRs from the communication network and from the NMR for the mobile device;
(c) applying the calculated weight to a location for each of said select ones of the NMRs from the communication network; and
(d) calculating a weighted sum of the locations.
4. The method of claim 1 wherein ones of the NMRs in the first set of NMRs are measured at different geographic locations in the communications network.
5. The method of claim 1 wherein the measurement data includes data selected from the group consisting of power levels, power differences, functions defined by power levels, functions defined by power differences, serving cell, serving sector, Timing Advance, Timing Advance equivalent, an ordered list of signal sources, round trip time, time difference of arrival, power density function of power levels, joint power density function of power levels, mobile device orientation, network state reported by the network, network state determined by an observing device, and combinations thereof.
6. The method of claim 1 wherein the observation cluster is determined as a function of a common parameter of the measurement data in the observation space.
7. The method of claim 1 wherein the common parameter is determined as a function of the following relationship:
d L1 2 ( i,j )= d O ( i,j )* d NS ( i,j )*( d RCV 2 ( i,j )+ d DIV 2 ( i,j ))
where i and j each represent an NMR, d O ( i,j ) represents a function comparing the observed measurements for ordered lists, d NS ( i,j ) represents a function comparing network states, d RCV 2 ( i,j ) represents a function comparing real continuous variables, and d DIV 2 ( i,j ) represents a function comparing discrete integer variables.
8. The method of claim 1 further comprising the step of determining a second observation cluster using a second set of NMRs, wherein the second set of NMRs includes at least the candidate NMR and NMRs from the first set.
9. A method of determining the geolocation of a mobile device operating in a region served by a communications network, the method comprising the steps of:
(a) determining, in observation space, an observation cluster of measurement data from a first set of network measurement reports (“NMRs”) from the communication network;
(b) comparing measurement data from a candidate NMR for a mobile device with the determined observation cluster of measurement data; and
(c) determining a location of the mobile device as a function of the compared measurement information,
wherein the step of determining an observation cluster further comprises the steps of:
(i) selecting an arbitrary NMR from the first set of NMRs;
(ii) extracting NMRs from the first set of NMRs having a common parameter with the arbitrary NMR;
(iii) generating a subset of NMRs including the extracted NMRs and the arbitrary NMR;
(iv) if the subset of NMRs contains fewer NMRs than a predetermined threshold, then iterating steps (i)-(iii) until the threshold is met; and
(v) if the subset of NMRs meets the predetermined threshold, then determining a focus of the subset of NMRs in the observation space.
10. The method of claim 9 wherein the step of comparing measurement data further comprises:
(i) assigning windows for each parameter type in the clustered measurement information;
(ii) comparing measurement information from the candidate NMR with the clustered measurement information from the subset of NMRs using assigned windows of a corresponding parameter type;
(iii) if the compared measurement information from the candidate NMR does not substantially correspond with a respective window, then adjusting the respective window until the compared measurement information from the candidate NMR substantially corresponds and re-clustering measurement information from a second set of NMRs; and
(iv) if the compared measurement information from the candidate NMR substantially corresponds with a respective window, then determining a location of the mobile device as a function of the location corresponding to the compared measurement information.
11. The method of claim 10 wherein the re-clustering further comprises:
(i) extracting measurement information from the second set of NMRs, the second set comprising at least the candidate NMR and NMRs from the subset or first set;
(ii) selecting an arbitrary NMR from the second set of NMRs;
(iii) extracting NMRs from the second set having a common parameter with the arbitrary NMR;
(iv) generating another subset of NMRs including the extracted NMRs and the arbitrary NMR;
(v) if the another subset of NMRs contains fewer NMRs than a second threshold, then iterating steps (ii)-(iv) until the second threshold is met; and
(vi) if the subset of NMRs meets the second threshold, then determining a focus of the subset of NMRs in the observation space.
12. A method of determining the geolocation of a mobile device operating in a region served by a communications network, the method comprising the steps of:
(a) determining, in observation space, an observation cluster of measurement data from a first set of network measurement reports (“NMRs”) from the communication network;
(b) comparing measurement data from a candidate NMR for a mobile device with the determined observation cluster of measurement data; and
(c) determining a location of the mobile device as a function of the compared measurement information,
wherein the step of comparing measurement data further comprises:
(i) assigning windows for each parameter type in the clustered measurement information;
(ii) comparing measurement information from the candidate NMR with the clustered measurement information from the first set of NMRs using assigned windows of a corresponding parameter type;
(iii) if the compared measurement information from the candidate NMR does not substantially correspond with a respective window, then adjusting the respective window until the compared measurement information from the candidate NMR substantially corresponds and re-clustering measurement information from a second set of NMRs; and
(iv) if the compared measurement information from the candidate NMR substantially corresponds with a respective window, then determining a location of the mobile device as a function of the location corresponding to the compared measurement information.
13. The method of claim 12 wherein the re-clustering further comprises:
(i) extracting measurement information from the second set of NMRs, the second set comprising at least the candidate NMR and NMRs from the first set;
(ii) selecting an arbitrary NMR from the second set of NMRs;
(iii) extracting NMRs from the second set having a common parameter with the arbitrary NMR;
(iv) generating a subset of NMRs including the extracted NMRs and the arbitrary NMR;
(v) if the subset of NMRs is less than a second threshold, then iterating steps (ii)-(iv) until the second threshold is met; and
(vi) if the subset of NMRs meets the second threshold, then determining a focus of the subset of NMRs in the observation space.