Predicting Fine-Grained Traffic Conditions via Spatio-Temporal LSTM

Table 3

Applications and comparisons about the predicted traffic conditions in navigation planning.

Literature	Raw data source	Object-based	Pathway	Core algorithm	Complexity	Accuracy	Save time	Perdurability
[1]	smart sensors	city	self-aware	Gaussian Process Regression	middle	middle	middle	low
[23]	GPS points	region	autonomous	Value Iteration Network	middle	middle	high	middle
[24]	street-view images	intersection	autonomous	CNN+RL+	high	middle	high	low
[25]	GPS points	region	agents	Ant Colony+RL	middle	middle	middle	middle
[26]	vehicles sharing	city	RIS	statistics	low	low	middle	low