Conclusion

It is clear that there is a desire for increased indoor positioning accuracy in modern technology.  Products like Amazon Go will most likely revolutionize retail purchasing.  However, Amazon Go is also telling of the challenges of indoor positioning.  It is the first large scale, commercial application of indoor positioning where clients interact directly with the positioning system.  However, it can only position users when they are in a controlled environment, like a checkout corridor.
These same challenges are highlighted in our survey of papers in this field.  Almost every paper we read makes note of the fact that environmental factors such as obstructions, interference, and multipath signals make calculating RSSI and fingerprints very challenging.  
Fluctuations in readings suggest that the best solutions to indoor positioning will take into account multipath mitigation or some RSSI filtering algorithm to reduce noise.  This includes environmental conditions as well.  Taking into account the factors that can cause RSSI values to fluctuate has proven to be beneficial to accuracy in our surveyed papers.
We have also seen that triangulation/trilateration seems to have the most accurate results, although fingerprinting seems to be the most heavily researched.  When triangulation or trilateration is researched, accuracy is nearly below 1 meter.  With this in mind, gaining further insights into how RSSI values can be fine tuned based on environmental conditions could increase accuracy as well. Additionally, using 3D triangulation could prove to be more accurate than 2D triangulation, as it reduces error when applied to 2D positioning.
When using fingerprinting as a positioning method, it seems to work better to use an average of decision making trees as your machine learning algorithm rather than others proposed.  In general, it seems that averages of values, whether in machine learning algorithms or collected RSSI's, helps to eliminate noise and provide more accurate positioning.
Based on our survey, these are the suggestions that we believe will contribute to more accuracy and further development in the field of indoor positioning.  Making indoor positioning a viable option for real world LBS's is a huge untapped market and research area.  By implementing these suggestions, we feel the future of indoor positioning will be realized sooner and more efficiently, while avoiding pitfalls that have been noticed in past research.