1 Introduction

Currently, most of the GPS devices adapted for the visually impaired people are used in urban areas. For instance: the breeze Human-Ware [1], Kapten Mobility Kapsys [2], GPS application from TeleorionVox [3], etc. In addition, last generation GPS devices are sensible to interferences and the heading is not precise at low speed (pedestrian navigation). If the user hesitates at low speed and loses his orientation, he has to walk with the GPS some tens of meters in order to have an updated heading from the GPS [4]. Additionally, most of the GPS devices are based on routes from the maps; however, in the case of the hiking roads, most of the routes do not exist or if exists they are not voice adapted for the visually impaired people. On the other hand, it is found some investigation in which it is used a voice based descriptive map in order to communicate geographic information with interactive map applications in hiking contexts for elderly people [5].

Nowadays, if a visually impaired person wants to practice hiking, a sighted person guides him. In this case, the visually impaired person holds on to the sighted person’s arm or shoulder as they walk along. If the road is narrow, the two persons have to walk in single file when moving.

In order to give the possibility, for the visually impaired people, to hike without sighted intervention, the authors have developed Navi Rando which is an adapted Smart Phone application that uses the information coming not only from the GPS, but also from IMU sensors (accelerometers, gyrometers) and a compass to give guidance instructions. Because of the use of the IMU sensors, the application is able to give the instructions even when the signal GPS is lost. The application has two parts, the application to record a new path or to modify a prerecorded path and the application to navigate.

2 Materials

The application to record a path, for making the path accessible to the visually impaired person, allows us to save every GPS coordinates in situ and to give comments related, for example, to the type of the ground (“take path with grassy ground”), or to anticipate and to give the information if there is any danger; for instance: “walk carefully, river on the right of the path”, or to clarify if there is any ambiguity in the path; for example, when there is a bifurcation, the comment can be “keep right” or “keep left” in order to take the good path. This information is associated to every GPS point recorded, if it is necessary.

The application to navigate uses the recorded path and gives the vocal information taking into account the body axis direction of the user in the form of “heading” and “distance” to the point, for example: “Point 5, 12 o’clock, 100 m, next point to 3 o’clock”, means that the blind person is walking already in the good direction to the GPS coordinates of the point 5 and he has to walk straight during 100 meters, once arriving to the point 5, to going to the point 6, he has to turn right and continue. This information can be given every 10, 15 or 20 s during the navigation. If the blind person is approaching to the point, the distance decreases, moreover, when the person arrives to the current point, the system automatically gives the information to the next point allowing the visually impaired person navigates from point to point. If the GPS signal is lost, the system will take into consideration the information from the IMU sensors in order to continue to give the guidance instructions.

In order to have the two hands free (one hand for using the white cane and the other one for using eventually a walking stick) and to listen the vocal information, the Smart Phone is put inside a pouch on the chest of the person (for example, light blue pouch, Fig. 1).

Fig. 1.
figure 1

Visually impaired person hiking using, on the right hand the white cane, on the left hand the hiking stick and the smart phone inside the light blue pouch. (Color figure online)

Full size image

3 Methodology

In order to run the tests, a sighted expert recorded the hiking path of 4.65 km close to the small town of Wingersheim in France with 30 GPS coordinates (see Fig. 2, on the left) giving comments if it was necessary (for example, at the round-about take the second route on the right, take the path with gravel ground, etc.).

Fig. 2.
figure 2

On the left, the cartography of the hiking road loop with 4.65 km. On the right, the conditions of the road (grassy ground with puddle).

Full size image

The path is a loop starting and finishing at a bus station in the small town. The urban area part has about 800 m crossing small streets and a roundabout, the rest of the path is rustic path having grassy ground with puddle (the day of the test was a rainy day) and gravel ground (see Fig. 2 on the right). Three visually impaired (two partially blind people and a fully blind person) testers between 20 and 33 years old from a blind association in Strasbourg agreed to be volunteers of this test.

These three visually impaired persons are experts in the use of the application, the instructions given (with more than 100 km hiking with the application in different kinds of grounds) and the use of the white cane for the hiking. None of the subjects had walked this hiking trail before. As, the hiking tests were done during the night, the first partially blind person (1st subject) still had residual vision during the night, the 2nd subject (partially blind person as well) had full night blindness and the 3rd subject is a fully blind person. Every visually impaired person started from the departure point every 5 min one after the other. The information: duration time, distance walked, GPS trajectory and average speed was collected during the tests with another application working at the same time of the navigation.

4 Results

The Table 1 shows the total distance of the trajectory walked and the time taken of the three visually impaired people. The Fig. 3 shows the GPS trajectory of the three visually impaired subjects.

Table 1. Total distance and time taken.
Full size table
Fig. 3.
figure 3

GPS trajectories of the three subjects (1st subject in cyan line, 2nd subject in red line and 3rd subject in green line). (Color figure online)

Full size image

We have also analyzed two parameters, the average speed of the three visually impaired in straight lines with no hesitation (Table 2) and the average speed every time that the subject has changed of direction. During the whole test, there have been thirteen changes of directions that have been called “turnings”. In order to calculate the average speed for the turnings, there were taken into account the speeds from 30 m before arriving to the turning until 30 m after passing the turning. The Table 2 shows the average speed in straight segments and Table 3, shows the rate between the average speed of turnings and the average speed of straight segments with every visually impaired person.

Table 2. Average speed in straight segments
Full size table
Table 3. Rate between the average speed of turnings and the average speed of straight segments; NA: Not Applicable.
Full size table

Let’s consider a good efficiency in turning, (considering the test conditions), when the person does not hesitate much to turn and the rate related to the average speed of the straight segments is more than 70 %. According to Table 3, the authors can point out that the first subject did 10 turnings (1st, 5th, 6th, 7th, 8th, 9th, 10th, 11th, 12th, and 13th) with good efficiency and the others with a slightly more hesitation. This can be explained because the 2nd turning is a roundabout and he paid more attention before crossing the streets making him reducing his speed because of the cars and the circulation. The 3rd and 4th were turnings segments with difficulty to walk and also to find with the white cane because of the ground with puddle making him reducing his speed as well.

For the second subject, it can be observed that he did 5 turnings (5th, 6th, 7th, 11th and 13th) with good efficiency, for the rest of the turnings; he had more hesitation being more critically the 3rd and the 9th turning. The decreasing in speed for the 1st and 2nd turning can be explained because he was in the urban area and he paid more attention to cross the streets and the roundabout. The 3rd turning, where the subject went out and deviated completely from the real path because he could not find the path with his white cane due to the rustic ground with puddle. For the 9th turning, he could not either find the path with his white cane and hesitated more, after he stopped for some time, went out the smart phone from the pouch and used the function “finding the current point permanently” which gives just the “heading” and “distance” to the current point and he could arrive to this point and came back to the real path for continuing with the navigation.

The third fully blind subject did 7 turnings (5th, 7th, 8th, 9th, 10th, 11th and 12th) with good efficiency; on the other hand, in the rest of the cases, the most critical turning was the 2nd one because he had to cross a roundabout in the urban area. The 3th and the 4th turning were not considered in the calculus because the person deviated completely from the real path passing for another path and arriving after directly to the 5th turning. During this time of deviation, he was walking slowly and hesitating trying to find a clear path to arrive to the next point.

5 Conclusions

Even in extreme conditions; for example, the rustic path was not clear enough to be detected with the white cane, mostly covered with puddle, the test was done in a rainy day (bad GPS reception conditions) and the hiking was during the night, the three visually impaired experts arrived to the final destination using the application without any form of sighted assistance and they were able to understand and get back from their errors.

The first partially blind subject with residual vision during the night did not go out from the path; he was just deviated slightly from the real path. He was the most efficient in doing the turnings compared to the other two subjects. This is explained because of his residual vision, even during the night, he could distinguish in some parts of the road the border of the path. This information in addition to the information coming from the navigation instructions and the white cane confirmed he was in the correct direction.

The second partially sighted subject with night blindness did 5 turnings with good efficiency out of 13, it is due to the fact that he has less experience in full blindness conditions for finding a path with the white cane in turnings. However he was faster in straight segments than the other two subjects. The third fully blind person walked slower but had a good efficiency in turnings compared to the second subject. In addition, he was able to recuperate the route after an alternate path.

We have tested the possibility to save a navigation path file and to use it to give instructions to guide the visually impaired people using Navi Rando in hiking trails. The efficiency of the guiding system seems to be accurate under the following conditions: firstly, the user has to be trained for the use of the white cane in hiking conditions, he needs also to be used to the guidance information “heading” and “distance” and he needs to find his own strategy to navigate using the information coming from the white cane to detect the path, from the GPS and from his own sensorial information.