Table A1.

Evaluation of reviewed systems based on addition features that caused that limitations of each system.

System Name/Weight/Type of Usage	Type of the Sensors	Accuracy	Analysis Type	Coverage	Measuring Angle	Cost	Limitation	Day/Night	Object Detection Range (Max/Min)	Classification Objects (Dynamic/Static)	Used Techniques for Detection, Recognition or Localization
Smart Cane Weight: N/A Type of usage: pilot stage	Ultrasonic sensors Water detector	N/A	Real time	Outdoor (only areas have RFID tags)	N/A	High	The water sensor can’t detect the water if it is less than 0.5 deep. The buzzer won’t stop before it is dry. A power supply meter reading needs to be installed to track the status	Day	1 m–1.5 m	Static	Ultrasonic technology
Eye Substitution Weight: light Type of usage: pilot stage	2 Ultrasonic Sensors Vibrator motors	N/A	Real time	Outdoor	Each sensor has a cone angle of 15°	$1790	The design of the system is uncomfortable due to the wood foundation which will be carried by the user most of the time as well as and the figures holes. The team used 3 motors for haptic feedback. They could use a 2-d array of such actuators that can give feedback about more details. Limited use by only Android devices	Day/Night	2 m–3 m	Static	GPS, GSM, and GPRS Ultrasonic technology
Fusion of Artificial Vision and GPS Weight: N/A Type of usage: deployment stage	Optical Sensors Bumble bee Stereo Camera 3-axis Accelerometers Electronic compass Pedometer	Accurate results for user position	Real time	Outdoor	6° of visual angle with (320 × 240 pixel) and 100° field of view with (640 × 480 pixel)	low	The system was tested on the function of the object’s avoidance technique. The system has not been tested or integrated with navigation systems to insure its performance; whether it will enhance the navigation systems as the authors promised or not is unknown.	Day	2 m–10 m	Static/dynamic	Global Position System (GPS), Modified Geographical Information System (GIS) and vision based positioning SpikNet was used as recognition algorithm
Banknote Recognition Weight: N/A Type of usage: pilot stage	iV-CAM	80%	Real time	N/A	N/A	low	This device was tested only on the Thai banknotes and coins, and it is not capable of working on other currencies that have similar colors of banknotes or similar sizes of coins. The device needs a method that controls the natural light that is used	Day	Closed View	Static	RGB model Banknotes Classification Algorithm
TED Weight: light Type of usage: pilot stage	Detective Camera	The corresponds are based on the feeling on the dorsal part of the tongue, (1,2,3,4) 100% (7) 10% (5,6,8) 50%	Real time	Outdoor	N/A	low	Antenna is not omni-directional. The range of voltage is not enough to supply the device. It is more difficult to recognize the pulses on the edges of the tongue.	Day/Night	N/A	Static	Tongue–Placed Electro tactile Display
CASBlip Weight: N/A Type of usage: pilot stage	3D CMOS sensor	80% in range of 0.5 m–5 m and less than 80% with further distance	Real time	Indoor/outdoor	64° in azimuth	N/A	Small detection range Image acquisition technique needs more than 1X64 CMOS image sensor. Acoustic module needs to be improved (it can add sounds in elevation)	Day/Night	0.5 m–5 m	Static	Binaural Acoustic module Multiple double short-time integration algorithms (MDSI)
RFIWS Weight: N/A Type of usage: research stage	None	N/A	Not-Real time	Outdoor	N/A	N/A	Collision of RFID Each tag needs specific rang which needs to be tested separated (scoop limitation) The tags cannot read the radio waves if case these tags get wrapped up or covered.	Day/Night	1 m–3 m	Static	Ultra-high frequency (UHF)
A Low Cost Outdoor Assistive Navigation System Weight: N/A Type of usage: pilot stage	3 Axial accelerometer sensors Magnetometer sensor	Good accuracy within residential area, but not as in an urban environment	Real time	Outdoor	N/A	$138	The accuracy of GPS receiver in high rise building is degraded. Limited scope, the GPS receiver needs to be connected via Bluetooth to perform.	Day	N/A	Static	GPS technology Geo-Coder-US Module MoNav ModuleBluetooth
ELC Weight: 0.170 Kg Type of usage: deployment stage	Ultrasonic sensor Micro-motor actuator	N/A	Real time	Outdoor	N/A	N/A	It is a detector device for physical obstacles above the waist line but the navigation still relies on the blind person.	Day/Night	Close objects over the waistline	Static	Ultrasonic sensor technology Haptics and tactile techniques
Cognitive Guidance System Weight: N/A Type of usage: pilot stage	Kinect sensor Video camera stereo Imaging sensor sonny ICx424 (640 × 480) RBG-D sensor for 3D point	N/A	Real time	Indoor	180°	N/A	Only 49 Fuzzy rules were covered which cover 80 different configurations. The perception capacities of the system need to be increased to detect spatial landmarks. Improve the stabilization of reconstructed walking plane and its registration through the frame.	Day	1.5 m–4.0 m	Static	The Canny filter for edge detection. Stereo vision, vanishing point and fuzzy rules (fuzzy logic and Mandani fuzzy decision system) to infer about the distances of objects.
Ultrasonic Cane as a Navigation Aid Weight: light Type of usage: pilot stage	Ultrasonic sensor (trans-receiver) Arduino UNO microcontroller wireless X-bee S1 trans receiver module	N/A	Real time	Indoor	30°	N/A	Just an object detector Small detection rang Does not detect objects that suddenly appear	Day/Night	5–150 cm	Static	Ultrasonic Technology
Obstacle Avoidance Using Auto-adaptive Thresholding Weight: N/A Type of usage: pilot stage	Kinect’s depth camera	N/A	Real time	Indoor	Horizontal 57.50° and Vertical 43.5°	N/A	The accuracy of Kinect depth image decreases when the distance between the scene and sensor increase. Auto-adaptive threshold could not differentiate between the floor and the object after 2500 mm. That increases the average error of distance detection. The depth camera has to be carried which is a lot of load on the user’s hand.	Day	0.8 m–4 m	Static/dynamic	Auto-adaptive Thresholding (divides equally a depth image into three areas. It finds the most optimal threshold value automatically (auto) and vary among each of those areas (adaptive).
Obstacle Avoidance Using Haptics and a Laser Rangefinder Weight: N/A Type of usage: pilot stage	Basely the system was built on the use of laser but the Novint Falcon has Encoder LED emitters and photo sensors Supplementary Sensors	N/A	Real time	Indoor	Horizontal 270° in front of chair	N/A	Precise location of obstacles and angles were difficult to determine.	Day	20 m with 3 cm error	Static	Haptics and a Laser Rangefinder
A Computer Vision System that Ensure the Autonomous Navigation Weight: N/A Type of usage: deployment stage	Monocular camera	High Accuracy	Real time	Indoor/outdoor	Angular field of camera view of 69°	low	Their fixed sizes of the image based on the category can make detecting the same object with different sizes a challenge. Since the proposed system is based on a smartphone video camera; if the video camera is covered by the blind person’s clothes, then the system cannot work. The objects are in dark places and highly dynamic objects cannot be detected. The overhead and noise of smartphones videos. The tested dataset of 4500 images and dictionary of 4000 words are considered as a small dataset. The system is tested and it works only on a Samsung S4 which makes it limited in scope.	Day	Up to 10 m	Static/Dynamic	Lucas–Kanade algorithm and RANSAC algorithm are used for detection. Adapted HOG descriptor extractor, BoVW vocabulary development and SVM training are used for recognition.
Silicon Eyes Weight: N/A Type of usage: research stage	24-bit color sensor SONAR obstacle detection light sensor 3-axis MEMS magnetometer 3-axis MEMS Accelerometer	N/A	Not-Real time	Not tested	N/A	N/A	A power supply meter reading needs to be installed to track the status. Low accuracy of GPS receiver in high rise buildings. The haptic feedback is not efficient. Limited memory of 2 GB micro-SD card to save user information.	Not tested	2.5 cm–3.5 m	Static	GPS & GSM technology
A Path Force Feedback Belt Weight: N/A Type of usage: research stage	IR sensor Two depth sensors (sensor 2 dual video cameras type Kinect)	N/A	Not-Real Time	Outdoor	360° over the blind’s waist	N/A	The detection range for this design is too small. The user needs to be trained in differentiating the vibration patterns for each cell. Using vibration patterns as feedback instead of audio format is not an excellent solution as the person can lose the sense of discrimination of such techniques over the time.	Not tested	Short	Static/dynamic	Infrared technology and GPS
EyeRing Weight: N/A Type of usage: pilot stage	Atmel 8 bit microcontroller OV7725 VGA CMOS sensor for image acquisition	N/A	Real time	Indoor/outdoor	Not Applicable	N/A	The system does not provide a real time video feedback. The system is limited to single object detection, which cannot be very useful to the disabled person.	Day	Close up view	Static	Roving Networks RN-42 Bluetooth module
FingerReader Weight: N/A Type of usage: pilot stage	Atmel 8 bit microcontroller OV7725 VGA CMOS sensor for image acquisition Vibration motors	93.9%	Real time tactile feedback20 m processing time	Indoor/outdoor	Not Applicable	N/A	There is a real time response for the audio feedback, but there is a long stop between the instructions. Also, the system prototype contains two pieces one is the ring, the other is the computation element which need to be carried all the time by the user for I/0 speech, otherwise the user will not be able to receive the feedback.	Day	Close up view	Static	Roving Networks RN-42 Bluetooth module
Navigation Assistance Using RGB-D Sensor With Range Expansion Weight: N/A Type of usage: pilot stage	RGB-D sensor	95%	Real time	Indoor	N/A	low	The effective of the infrared to the sunlight can negatively affect the performance of the system outdoors and during the day time.	Night	Up to 3 m using range information technique and from 3 m and further using the vision information	Static	RANdom Sample Consensus (RANSA) detection algorithm Image intensities and depth information (computer vision) Infrared technology and density images
Mobile Crowd Assisted Navigation for the Visually-impaired Weight: N/A Type of usage: pilot stage	Camera GPS Compass Accelerometer	20.5% improvement in crowd sound for navigation	Real time	Indoor	N/A	N/A	The collected information is based on the volunteers’ availability. There is a possibility of no input in the interval time which fails the goal of the service.	Day/Night	N/A	Dynamic	Crowd sounding service through Goagle engine for navigation Machine vision algorithm
A Design of Blind-guide Crutch Based on Multi-sensors Weight: N/A Type of usage: deployment stage	3 Ultrasonic sensors	N/A	Real time	Outdoor	30° detection range for 2 sensors, 80° detection range for overhead	N/A	The detection range is small. This system is claimed to be navigation system, however, there are no given directions to the user.	Day	0 m–2 m in front	Static	Ultrasonic distance measurement approach
Ultrasonic Assistive Headset for visually-impaired people Weight: light Type of usage: pilot stage	4 Ultrasonic type (DYP-ME007) sensor obstacle detector	N/A	Real time	Indoor/outdoor	60° between ultrasonic distance sensors	N/A	Limited directions are provided. The headset obscures the external noise.	Day/Night	3 cm–4 m	Static	Ultrasonic technology
A Mobility Device for the Blind with Improved Vertical Resolution Using Dynamic Vision Sensors Weight: light Type of usage: pilot stage	2 retine-inspired dynamic vision sensors (DVS)	99% ± object detection, 90% ± 8% horizontal localization, 96% ± 5.3% size discrimination	Real time	Indoor	N/A	low	The modules are very expensive. Further intensive tests need to be done to show the performance object avoidance and navigation techniques, whereas, the test was mainly on object detection technique for the central area of the scene.	Day	0.5 m–8 m	Dynamic/static	Event-based algorithm
Ultrasonic for ObstDetectRec Weight: 750 gram Type of usage: pilot stage	4 ultrasonic sensors (Maxsonar LV EZ-0)	N/A	Real time	Indoor/outdoor	±40°	Low	The system cannot detect obstacles above waist level. There is no navigational information provided. Small detection range. It is not an independent device.	Day	2 < R ≤ 5 m	Static/dynamic	Vision-based object detection module. Ultrasonic technology. SVM
SUGAR system Weight: N/A Type of usage: pilot stage	Ultra-wide band Sensors(UWB)	High Accuracy	Real time	Indoor	N/A	N/A	Sensors would have to be deployed in every room. The room has to be mapped beforehand. User needs to select destination beforehand. It is not suitable for outside use.	Day (the system was not tested for night time)	50 m–60 m	static	UWB positioning technique Path Finding Algorithm Time Difference of Arrival technique (TDOA)

* Not Available online: N/A.