1. Introduction
When advancing in age, older people may face difficulties walking in the city due to physical, but also, perceptual and cognitive declines [
1]. Finding one’s way, searching for relevant information in the environment, and crossing streets while being aware of incoming cars are some examples of problems older pedestrians may face [
2]. These difficulties can lead the older people to reduce the frequency of their urban travels and can have consequences on their wellbeing and quality of life [
3], as pedestrian mobility is one of the main means of daily transportation for older adults in French cities [
4]. Spatial navigation, more particularly, may be a source of trouble for older pedestrians. It consists in finding and traveling a route from a point A to a point B [
5,
6] and involves complex cognitive and perceptual processes such as self-orientation [
7], attention to the environment and selection of landmarks [
8], use and interpretation of allocentric representations [
9], construction of cognitive maps [
10], or maintenance of the goal in mind to reach the destination [
11]. However, all these processes tend to decline with aging, leading to greater difficulties in navigation among older people.
Paper and digital maps are the media most commonly used to help people with navigation tasks [
12,
13]. However, maps are not very suitable for older pedestrians, even if they are used to them. Indeed, they provide people with allocentric representations of the environment that are difficult to interpret for older people, who prefer egocentric (first-person view) types of information to help them find their way [
9,
14]. Maps are also difficult to orient correctly [
15] and require a lot of attention to be read [
16,
17], leading to potentially dangerous situations for older people whose attention resources are limited [
18,
19]. To guide them in the city, one solution could be to provide older pedestrians with turn-by-turn egocentric instructions that are simple to understand (e.g., turn right at the next intersection [
14]). To do so, wearable devices could be an interesting option as they could help reducing attention sharing by providing relevant information directly on the pedestrian’s body. Moreover, wearable devices allow using several sensory modalities to help people perceive navigation instructions, thereby triggering potential benefits for older people whose sensory acuity may decline, while the maps only rely on the visual channel.
In this paper, we investigate the performance and user experience of older pedestrians with four navigation aids used to find their way in the city: (1) the paper or digital map, i.e., an aid people usually use in such circumstances and three sensory wearable devices: (2) augmented-reality glasses (visual instructions); (3) bone conduction headphones (auditory instructions); and (4) a smartwatch (warning haptic vibrations + visual instructions). The main issues raised in this study are as follows. Would these devices be more efficient than a traditional paper map? Would they be well accepted by older pedestrians? Are some of these devices more efficient and appreciated?
5. Discussion
This study aimed to compare the navigation experiences (in terms of efficiency, user experience, and inattentions) of older pedestrians using their usual navigation aid (paper and digital maps) and three sensory wearable devices to find their way in an urban environment. The sensory navigation aids were based on market-available devices (AR glasses, bone conduction headphones, and a smartwatch) and turn-by-turn navigation instructions were designed using previous literature among older people (see e.g., [
30]).
Most of the participants (16 out of 18) used a paper map when asked to bring their own navigation aid. This observation was in-line with previous studies [
27,
30] with older people using preferentially paper maps as pedestrians, while GPS have become usual among older drivers.
The results highlighted that the participants took a longer time to reach their destination with their usual aid than with the three wearable devices. This could be explained by the difficulty older pedestrians usually face when navigating an outdoor environment with a map [
15,
19]. The time needed to prepare the route with the map, in particular, differed a lot among the participants. Preparing the route is a key element for older people in unknown areas [
74], and the high interindividual variability among participants may reflect important differences in the participants’ trust in their orientation capabilities [
75]. Some studies previously brought into light that older pedestrians varied considerably in the self-evaluation of their spatial abilities [
30,
76].
The percentage of success in turning into the right street was higher with the bone conduction headphones, the smartwatch, and the usual aid than with the AR glasses. In some cases, the percentage of success was even 100% (with the smartwatch in roundabouts), thereby confirming that technological aids could help compensate the difficulties older pedestrians usually face in urban environments [
61]. The percentage of success in roundabouts was close to the percentage of success in simple intersections, although we might have supposed that the difficulty would have been higher in roundabouts [
61]. The participants were often more attentive to the instructions in the roundabouts than in the simple intersections because they understood well the functioning of the system in this type of configuration (the use of two instructions in these areas) and were able to anticipate the second instruction to exit to the right street. However, the way people subjectively perceive the environment may differ from that of the geographic databases used by the navigation system. For example, some configurations were labeled as roundabouts by the participants but not by the system, which led to a discrepancy between participants’ expectancies and the system’s functioning. Thus, using geographic databases adapted to the pedestrians’ navigation would be necessary instead of using databases for cars’ navigation, pedestrians having a different perspective than drivers, and more possible paths [
27].
The percentage of success in finding the right way was lower with the AR glasses but it did not differ between the bone conduction headphones, the smartwatch and the usual aid of the participants. The average percentage of errors was low among the participants using their own map, which could be explained by the familiarity with this aid and the longer time required to prepare the route before leaving.
Even though the performance did not differ between the bone conduction headphones, the smartwatch, and the usual aid, differences were observed in the way the participants acted to find their way. Using the usual aid seemed to rely on a sum of spatial cues found from both the map (paper or digital) and the environment (topography, street names, etc.). The participants actively searched for these clues to solve the spatial problem they were facing. Hence, the participants were free to choose the clues needed for their navigation at their own pace and thus were possibly more autonomous than with the wearable devices. Searching for clues in the environment, however, requires a high level of attention and it generated many instances of inattention, as was highlighted by the videos’ analysis, even if the participants do not seem conscious of these episodes with their usual aid during the interviews. This highlights a potential benefit of wearable navigation aids in terms of safety with which the number of instances of inattention was low. An active search for clues also requires spatial cognitive abilities, which vary a lot among older people and could be specifically altered [
8].
Navigation with turn-by-turn wearable devices relied on a more localized spatial environment (i.e., participants did not have to locate themselves on a map, just to look at the next street to turn in) and was more contingent to the trust the participants had in the navigation aid, as the trust in the device depended a lot on its temporality and its steadiness over time. Even though the participants actively solved spatial problems with the wearable devices too, the matching between the spatial clues found in the environment and the instructions provided by the system depended a lot on the system itself (i.e., the possibility to replay the instructions, the temporality of the instructions, and the rules of the system functioning). This could have led to a higher dependence to the system when navigating with such devices. Noticeably, the participants had the possibility to use a map with the three wearable devices, which they did not, highlighting that the navigation strategies were different with the wearable devices on one hand and the usual aid on the other hand.
The wayfinding processes based on the use of wearable devices such as the ones we developed for the present study purposes seemed to rely, like the usual map, on cognitive abilities (to understand the navigation instructions and mapping it to the environment), but also on other types of abilities related to the use of digital and dynamic artefacts such as memorizing the instructions, interpreting the interface, or understanding and inferring rules from the use of the system [
77]. For example, based on the previous instructions they had perceived, participants were able to infer the temporality of the instructions, and knew that the next instructions would be provided 15 to 20 meters before the intersection, otherwise they could go straight forward. In contrast, with their usual map, they tended to count the number of intersections before the next turn. In this way, wearable devices could thereby benefit older people with lower spatial cognitive capacities because using such devices rely on other types of abilities that directly depend on the system design.
Beyond poorer performance, the AR glasses also lead to a poorer user experience than the bone conduction headphones and the smartwatch, whatever the UX dimension. While perceiving and understanding the arrows instructions during the indoor familiarization phase was rather easy to the participants, the perceptibility of the instructions and their temporality in an urban navigation context was problematic, as well as the comfort and discretion of the sensory wearable device.
In our analyses, the user experience was shaped by three main phenomena over time. These phenomena are rather congruent with the three layers of need proposed by Fang et al. [
78] when using pedestrian navigation aids. Shifts in attention corresponds to the physical sense layer (the perception of visual, auditory, or haptic clues to find one’s way), understanding the situation corresponds to the safety layer (the ability to manage the risk and errors related to the route and the environment), and the affective and aesthetics feelings are related to the mental satisfaction layer (the need for confidence, comfort, and respect).
Shifts in attention: The arrows inlayed in the visual field by AR glasses and the warning vibration of the smartwatch were difficult to perceive for some participants. We have two possible explanations here: (1) this difficult perception is either due to the fact that the visual attention to the surroundings interfere with the visual perception of the arrows superimposed in the same field of view or (2) it may be due to an overly weak intensity and/or an impression of short duration when participants were actively involved in a navigation activity in an urban context. (We note the difference between the perceptibility indoor in a quiet place and in the urban context outdoor.) These results about the importance of perceptibility were already pointed out by [
49]. As the perceptibility of the arrows was reduced in the urban navigation context, participants were required to be more attentive to search for and perceive the navigation instructions. This had some consequences on the possibilities for participants to look at the environment and for ensuring their safety. Various strategies were used to improve the perception of the instructions, such as replaying instructions (the number of replayed instructions being higher with the AR glasses than with the two other devices) relying on preparative attention [
79], by looking precisely at the AR screen, or immobilizing the arm close to the chest with the smartwatch. Attention sharing also had some consequences on the feeling of safety with AR glasses. The participants tended to be more cautious because they feared they might not be at their best. We also observed that the participants tended to act more promptly, and, maybe, more automatically and instinctively, when being guided by a turn-by-turn device, which could have impacted their safety negatively.
Understanding the situation: We observed more doubts about where to go with the usual map and with the glasses. Contrary to the paper maps that were read at the participant’s own pace, the temporality of instructions (the moment the instruction was provided in comparison with the moment the participant would turn) seemed to be a key element for decision making with the wearable navigation devices. Impressions of unsteady and belated instructions led to difficulty in understanding the situation and had a negative impact on the participants’ trust in the aid, especially with the AR glasses. As previously mentioned, the matching between the spatial clues found in the environment and the instructions depended a lot on the system itself when using a wearable aid (the accuracy of the U-turn instructions, the replay button, the instructions temporality, and the system rules). Providing instructions 10 meters before the intersection appeared to be mandatory to ensure quiet and safe conditions for decision making.
Several elements could impact the temporality of instructions, among which the GPS accuracy and the quality of the connection between the smartphone and the wearable device. Despite poorer GPS accuracy in urban environments [
60] and 10-meter-precision of the GPS on smartphones [
80], these could not explain the differences subjectively perceived between the AR glasses and the two other devices we used. It could be supposed that the Bluetooth connectivity was poorer with the AR glasses than with the bone conduction headphones and the smartwatch, but we were not able to verify this hypothesis. The higher fear to miss a message and the poorer feeling of safety of the participants with the AR glasses were also likely to impact the perception of the temporality of instructions, as stress could constitute a bias for the perceived duration of stimuli [
81].
Affective and aesthetic feelings: the emotions were highly contrasted among the participants during the four routes with the four aids. Some of them were frustrated with not perceiving the instructions, while others were pleasantly surprised by their ability to navigate by themselves with a highly technological device. We observed that comfort and discretion were key concerns for the participants, with the bone conduction headphones and the smartwatch being more positively evaluated than the AR glasses. The AR glasses were described as being heavy on the nose and to lack discretion. Participants wished to be able to hide the device they were using, to look “like everybody” in the street. The aesthetics, discretion (e.g., color and size) and the weight of devices designed to be “wearable” should be considered carefully to boost the physical and psychological comfort among older pedestrians. Aesthetics is sometimes told to be a minor dimension in UX for older people [
57,
82], who are supposed to focus primarily on the pragmatic qualities of the devices rather than the hedonic ones. In this study, aesthetics was a major concern. In addition, it should be noted that adding sunshades to the AR glasses may have had consequences on the UX by reducing their discretion, even though it aimed primarily to improve the perceptibility of the arrows.
Understandability, perceived usefulness [
83], pleasantness [
84], and aesthetics [
47] are traditionally identified as elements that impact the user experience. Thus, the results of this study are consistent with the literature. However, we note that many dimensions were specific to older people’s user experience with navigation aids. The perceptibility of instructions and the related increase of attentional demand [
49], the temporality of instructions, and the accuracy of the GPS signal [
50] were indeed discussed in depth in the interviews. Also, the feeling of trust, the feeling of safety, and discretion are often associated with older people’s user experience [
52]. The study did not bring into light new dimensions of UX but helped understand how user experience was shaped by all these dimensions over time. Hence, the perceptibility of the instructions and attentional demand impacted the feeling of safety, which is paramount for older people. The older participants were not “passively” following the turn-by-turn instructions provided by the wearable devices, they had to at least understand the messages, the mode of operation of the system, and to keep an eye on the environment, thereby challenging the hypothesis of “infantilization” that is traditionally associated to the use of turn-by-turn aids [
23].
While our results highlight that navigation among older pedestrians could be fostered by the use of some wearable devices providing turn-by-turn instructions in comparison to a map, some limitations of our work should be acknowledged. The limited number of participants (18), due to the need to interview them in detail (each participant spent 4 hours in the study) could also be a limit to the generalization of the results. Also, we did not compare the performance and user experience of the older participants to younger ones. Similarly, we were not able to compare performance and user experience between men and women. The route and the interview were slightly different between the usual aid and the three wearable devices (but we observe that the simpler route for the usual aid does not benefit to its efficiency), and we were not able to compare some dimensions between the four conditions (such as aesthetics, as no participant mentioned it with the usual map); a total similarity of the conditions should be preferred in future studies.