. 2022 May 31;22(11):4208. doi: 10.3390/s22114208

Table 4.

Radar datasets.

Name	Year	Task	Radar Type	Data	Doppler	Range	Other Sensors	Scenarios	Weather	Annotations	Size
Radar Datasets for Detection
nuScenes [34]	2020	DT	LR	PC	✓	SV	CLO	USH	✓	3D, T	L
PixSet [35]	2021	DT	LR	PC	✓	MR	CLO	USP	✓	3D, T	M
RadarScenes [36]	2021	DTS	HR	PC	✓	SV	CO	USHT	✓	P $_{w}$	L
Pointillism [37]	2020	D	2LR	PC	✓	MR	CL	U	✓	3D	M
Zendar [38]	2020	D	SAR	ADC, RD, PC	✓	MR	CLO	U	✗	P $_{w}$	S
Dense [43]	2020	D	LR	PC	✓	LR	CLO	USHT	✓	3D	L
RADIATE [44]	2020	LDT	SP	RA	✗	SV	CLO	USHP	✓	2D, T, P $_{s}$	M
Radar Pre-CFAR Datasets for Detection
CARRADA [45]	2020	DTS	LR	RAD	✓	SR	C	R	✗	2D, P $_{w}$ , M, T	M
RADDet [46]	2021	D	LR	RAD	✓	SR	C	US	✗	2D	M
CRUW [47]	2021	D	LR	RAD	✓	USR	C	USHP	✗	P $_{o}$	L
RaDICaL [48]	2021	L	LR	ADC	✓	USR, SR	C $_{d}$ O	USHIP	✗	2D	L
Ghent VRU [49]	2020	DS	LR	RAD	✓	SR	CL	U	✗	M	M
4D Radar Datasets for Detection
Astyx [2]	2019	D	HR	PC	✓	MR	CL	SH	✗	3D	S
View-of-Delft [50]	2022	DT	HR	PC	✓	SR	CLO	U	✗	3D, T	S
RADIal [51]	2021	DS	HR	ADC, RAD, PC	✓	MR	CLO	USH	✗	P $_{o}$ , M	M
TJ4DRadSet [52]	2022	DT	HR	PC	✓	LR	CLO	U	✗	3D, T	M
Radar Datasets for Localisation
Oxford [53]	2020	L	SP	RA	✗	SV	CLO	U	✓	P $_{s}$	L
Mulran [54]	2020	L	SP	RA	✗	SV	LO	US	✗	P $_{s}$	M
Boreas [55]	2022	LD	SP	RA	✗	SV	CLO	S	✓	P $_{s}$ , 3D	L
EU Long-term [56]	2020	L	LR	PC	✓	LR	CLO	U	✓	P $_{s}$	M
Endeavour [57]	2021	L	LR	PC	✓	5LR	LO	S	✗	P $_{s}$	M
ColoRadar [58]	2021	L	HR, LR	ADC, PC	✓	2USR	LO	SIT	✓	P $_{s}$	M
Radar Datasets for Other Tasks
PREVENTION [59]	2019	DT	LR	PC	✓	1LR, 2SR	CLO	UH	✓	2D, T	L
SCORP [60]	2020	S	LR	ADC, RAD	✓	USR	C	P	✗	M	S
Ghost [61]	2021	DS	LR	PC	✗	LR	CLO	S	✗	P $_{w}$	M

Task: D, T, L, and S stand for detection, tracking, localisation, and segmentation; Type: LR, HR, SP, and SAR stand for low-resolution, high-resolution, spinning, and SAR; Range: SV, LR, MR, SR, an USR stand for surrounding view, long-range (<250 m), middle-range (<180 m), short-range (<50 m), and ultra-short-range (<25 m); Other Sensors: C, C_d, L, and O stand for camera, RGBD camera, LiDAR, and odometry; Scenarios: U, S, H, P, T, R, and I stand for urban (city), suburban, highway, parking lot, tunnel, race track, and indoors; Size: L, M, and S stand for large, medium, and small;Weather stands for adverse weather; Label: 2D, 3D, T, P_w, P_o, P_s, and M stand for 2D bounding box, 3D bounding box, track ID, pointwise detection, object-level point, pose, and segmentation mask.