Detecting incorrectly placed access points

A method, comprising:

• receiving a position map specifying a respective position for each of a plurality of access points within a physical environment;
• collecting indications of signal strength from each of the plurality of access points;
• and
• analyzing, by operation of one or more computer processors, the indications of signal strength and the position map to determine that a position of a first one of the plurality of access points within the position map is likely to be incorrect, in that the position of the first access point within the position map does not correspond to a position of the first access point within the physical environment, comprising:
• plotting a line of best fit using a path loss model;
• determining a likelihood value for each of the plurality of access points, by determining a respective average negative log likelihood value based on the line of best fit and the indication of signal strength and a calculated distance for the access point;
• and
• selecting an access point having the greatest average negative log likelihood value.

A non-transitory computer-readable medium containing computer program code that, when executed by operation of one or more computer processors, performs an operation comprising:

• generating a set of signal strength and distance data, comprising, for each access point in a plurality of access points:
• identifying a subset of the plurality of access points that neighbor the access point;
• collecting indications of signal strength for each of the access points in the subset, relative to the access point;
• calculate a distance between the access point and each of the access points in the subset, based on a position map;
• generating a line of best fit to the set of signal strength and distance data for each of the plurality of access points;
• determining a likelihood value for each of the plurality of access points, by determining a respective average negative log likelihood value based on the line of best fit, the indications of signal strength and the calculated distance for the access point;
• and
• determining the position within the position map for a first one of the plurality of access points is likely to be incorrect, by selecting an access point having the greatest average negative log likelihood value based on the determined likelihood values for each of the plurality of access points.