Day by day, aged 65 and above global population is increasing rapidly, and up to 2050, it will reach 1.5 billion. So, the fall detection system is a trending research area where the research community is showing interest day by day. Research in fall detection has seen an upward movement in the recent past. Falls to the elderly can be life-threatening or might cause severe injuries if the person remains on the ground for a longer duration without proper medical care. This paper summarizes comprehensive research that has taken place in the last decade to help new researchers in this field. These articles are taken from primary sources, and the selection criteria are mentioned. This paper summarizes each article's various algorithms, input parameters, classification, and limitations. The performance criteria and publically available datasets were analyzed to access the performance results. At last, we found that the field is slowly but constantly moving towards its implementation in Machine learning, CNN, and Deep learning. There is a lack of well-organized and real-world data to train the classifiers in the event of an actual fall of an older person

artificial vision; neural networks; fall detection; fall classification; fall dataset, CNN, Deep Learning