Air Combat Maneuver Trajectory Prediction Model of Target Based on Chaotic Theory and IGA-VNN

Table 4

Target Z-coordinate multistep prediction performance comparison of four different prediction methods.

Prediction step	Algorithm	RMSE	MAE	MAPE	COR
1	BP	3.8168	1.8070		0.9959
	IGA-VNN	1.4914	1.0943		0.9984
	GA-VNN	3.5207	1.7437		0.9958
	Volterra	2.0636	1.3294		0.9937
	RBF	5.0411	2.0680	0.0011	0.9720
2	BP	7.6877	2.6837	0.0019	0.9881
	IGA-VNN	4.9893	2.1206	0.0012	0.9968
	GA-VNN	4.9718	2.2086	0.0013	0.9957
	Volterra	6.2689	2.4025	0.0015	0.9479
	RBF	7.9969	2.3284	0.0018	0.9572
4	BP	11.9821	3.2566	0.0028	0.8768
	IGA-VNN	6.4740	2.3860	0.0015	0.9757
	GA-VNN	8.0413	2.7889	0.0021	0.9978
	Volterra	8.1816	2.5405	0.0017	0.9316
	RBF	14.0847	3.396	0.0032	0.9044
6	BP	14.1209	3.2762	0.0029	0.8273
	IGA-VNN	9.8469	2.8852	0.0022	0.9677
	GA-VNN	10.9808	2.9773	0.0024	0.9585
	Volterra	10.1371	2.9463	0.0023	0.8627
	RBF	20.9611	4.4327	0.0052	0.8633