Bridging the Gap: Using Mobile Augmented Reality to Reconnect Museum Artifacts at Lofotr Viking Museum in Norway with Their Original Contexts

3.1. The Borg Site: Archaeological Context and Interpretive Challenges

The Lofotr Viking Museum is located at Borg in Lofoten, on the site of an extensive Iron Age and Viking Age settlement. While key archaeological discoveries at Borg led to the establishment of the museum in the mid-1990s, much of the original cultural landscape is no longer directly visible to visitors. Structural remains such as house foundations, postholes, and burial mounds are largely absent or difficult to discern in the terrain, limiting visitors’ ability to perceive how artefacts relate spatially to their original settings.

To compensate for this loss of visible context, the museum combines indoor exhibitions with reconstructed architecture in the surrounding landscape. The reconstructed Viking chieftain’s house provides a strong experiential anchor; however, its placement adjacent to, rather than directly on, the original footprint introduces a spatial displacement that can complicate interpretation. Visitors may encounter the reconstructed building as if it were “the” original, while the original footprint and related archaeological traces remain difficult to perceive. This creates an interpretive tension between visible reconstruction and largely invisible archaeology.

Previous AR studies in comparable heritage settings indicate that users often struggle to understand how virtual reconstructions relate to physical remains when multiple representational layers coexist (Bernardini et al. 2012; Sierra et al. 2024). At Borg, this challenge is amplified by the close proximity of physical reconstructions to the original archaeological remains, creating conditions in which digital overlays must compete simultaneously with both visible structures and absent contexts.

In this hybrid environment, the relationships between artefacts, their original locations, and their subsequent reconstruction become difficult to apprehend without additional interpretive support. The present study, therefore, examines whether a site-wide mobile AR narrative can help users more clearly perceive object–place relationships at Borg, where physical reconstruction, archaeological absence, and digital simulation intersect.

3.2. Mobile AR and the Reconnection of Objects and Landscapes

Mobile AR uses smartphones or tablets to overlay digital layers, such as 3D models, text, audio, and video, onto the live camera view, guided by device sensors including GPS. Unlike static or viewpoint-fixed implementations often described as indirect AR (Wither, Tsai & Azuma 2011: 815), mobile AR supports dynamic, spatially aligned interaction in which bodily movement through the environment corresponds directly to movement within the simulation. In this study, the term mobile AR is used to emphasise real-time, sensor-driven synchronisation between user movement and digital content, rather than fixed or panoramic viewpoints.

At the Lofotr Viking Museum, this approach is implemented through a mobile AR prototype that enables visitors to explore reconstructed buildings anchored to their archaeological footprints in the terrain. SitSim Core, developed as part of the SitSimLab project (Liestøl n.d.), places 3D models of Viking and early medieval buildings at full scale in historically informed locations. As visitors move through the landscape, digital structures appear in relation to their physical surroundings, allowing to explore spatial relationships that are no longer legible in the present-day terrain.

The prototype also digitally re-places selected museum artefacts within these reconstructed environments. Artefacts are presented as interactive 3D objects that can be enlarged, rotated, and examined in detail, enabling users to consider their material characteristics and their relationships to architecture and activity areas. In this way, objects encountered indoors in the exhibition are recontextualised within plausible outdoor settings of use.

Interaction is structured through spatially anchored points in the landscape that function as hypertext links, activating audio, short texts, video, and interactive objects within a first-person camera view. The system further supports switching between alternative architectural reconstructions for selected features, such as roof materials or pitch angles. Rather than presenting a single authoritative model, these alternatives are intended to communicate archaeological uncertainty and ongoing scholarly debate.

Through this combination of movement-based interaction, spatial anchoring, and multimodal narrative cues, mobile AR reconnects fragmented archaeological traces with museum artefacts by situating interpretation directly in the landscape. Instead of treating reconstructions as detached visualisations, the prototype integrates scale, proximity, and bodily movement as interpretive resources, allowing visitors to encounter objects and structures in relation to their original locations.

Figures 2, 3, 4 illustrate how the architectural reconstructions were iteratively refined across different prototype versions. In the first version, Viking Age and medieval buildings were presented with limited visual differentiation, which several participants described as showing little seasonal variation and therefore making temporal change difficult to interpret. In the second version, adjustments to lighting and a shift in seasonal setting from summer to winter were introduced to more clearly emphasise temporal change.

Figure 2

Figure 3

Figure 4

4.1. Research approach: iterative design and in-situ evaluation

The study adopts a media design approach focused on the iterative development, testing, and refinement of a mobile AR prototype in a real-world museum setting. Rather than treating the prototype as a final product, media design frames it as a research instrument through which interpretive challenges and design decisions can be examined empirically.

From a methodological perspective, media design aligns with design science research, which emphasises the creation and evaluation of artefacts intended to address practical problems while simultaneously generating knowledge (Hevner et al. 2004). Within media and communication studies, Anders Fagerjord (2012) conceptualises media design as a scientific research method in which the systematic development and refinement of media artefacts constitute a form of scholarly inquiry. By treating design processes as analytical resources rather than merely production activities, his framework integrates theoretical reflection, critical evaluation, and experimental practice, positioning media artefacts, understood as concrete, produced media objects that can be analysed, tested, and reflected upon within the research process, such as a mobile AR prototype in the present study – as epistemic instruments in media research.

Also, within this tradition, Liestøl (2013) demonstrates how experimental prototypes, particularly mobile AR, can function as research instruments for exploring new forms of spatial narration and embodied interpretation. Central to this approach is iterative prototyping combined with testing in authentic contexts, enabling systematic refinement based on user interaction. Lars Nyre (2014) similarly conceptualises media design as a research method grounded in the development, testing, and evaluation of experimental media prototypes, with attention to both technological innovation and societal relevance.

While Liestøl, Fagerjord, and Nyre all conceptualise design as a mode of inquiry rather than mere production, the present study extends this tradition by embedding design experimentation within a museological and archaeological framework. This framework foregrounds empirical in-situ user testing, ethical transparency, and the analysis of embodied audience experience in heritage environments.

In this study, the approach is operationalised through four iterative stages: (1) creation of digital media linking museum artefacts to their original archaeological contexts; (2) in-situ user testing to assess usability and interpretive clarity; (3) identification of errors, ambiguities, and misunderstandings; and (4) modification of content, interface, and spatial alignment based on empirical findings. Two test rounds conducted under differing conditions enabled iterative comparison and supported the systematic refinement of the prototype, as illustrated in Figure 5.

Figure 5

4.2. Participants and Demographics

Two in-situ test rounds were conducted at the Lofotr Viking Museum, involving a total of 31 participants (January 2025: n = 12; September 2025: n = 19). Participants were recruited on site through convenience sampling and included both museum visitors and staff with varying degrees of familiarity with the museum and its landscape. The study aimed to gather formative feedback to inform iterative design development rather than to produce statistically generalisable results.

Participation was voluntary and based on informed consent. No personally identifying information was collected or reported. Age was recorded in both test rounds, while educational background was recorded in the second round only.

Participants in the first test round ranged in age from 20 to 89 years, with the largest groups in the 40–49 and 60–69 age categories. The second test round primarily included participants aged 20–44 years, with a smaller number aged 45–69 years. Educational backgrounds in the second round spanned a broad range, including high school education (n = 6), bachelor’s degree (n = 7), master’s degree (n = 3), doctoral degree (n = 1), and other relevant educational or professional backgrounds (n = 2). No participants reported primary school as their highest completed education. Age distributions across test rounds are presented in Table 1.

4.3. Procedure

Each test followed a predefined route through the museum grounds, beginning at the museum building and continuing through reconstructed structures and archaeologically relevant outdoor areas. Participants used the mobile AR prototype in situ under authentic environmental conditions, including snow and low temperatures in January and bright sunlight in September.

During the experience, participants encountered reconstructed buildings and spatially anchored digital content designed to illustrate relationships between exhibition objects and their original locations. After completing the route, participants filled out a post-use survey. Throughout the experience, go-along observation was conducted to capture navigation behaviour, interaction patterns, and moments of hesitation or confusion that might not be evident from survey responses alone.

4.4. Survey instrument

The post-use survey was designed to collect formative feedback on usability and interpretive clarity, with particular emphasis on participants’ understanding of object–place relationships. Responses were recorded using five-point Likert scales ranging from strong disagreement to strong agreement, supplemented by optional open-ended comments (Tanujaya et al. 2022).

The first test round employed 15 Likert-scale items, while the second round expanded the instrument to 30 items and four open-ended questions. Selected items were retained across rounds to enable comparison between iterations. A central item repeated in both rounds addressed the core research aim:

In addition, the first test round included an item assessing the perceived interpretive importance of object–place relationships, while the second round revised the wording to focus more explicitly on participants’ understanding after using the prototype. This revision reflected changes to the design and a shift from assessing general attitudes to evaluating experiential outcomes.

Survey data were analysed descriptively by examining distributions across response categories. Open-ended responses were reviewed in relation to quantitative patterns to identify recurring usability and interpretation issues. Comparison across test rounds made it possible to assess the impact of specific design modifications, including the introduction of explicit audio cues and the replacement of manually modelled objects with photogrammetric scans.

The full survey instruments included additional items addressing usability, engagement, authenticity, and learning outcomes; Table 2 presents selected items most directly related to the study’s research questions.

4.5. Observation

Observation was conducted during both test runs to capture aspects of user behaviour that could not be reliably inferred from self-reported survey data alone. A go-along approach was employed (Tjora 2021), in which participants were accompanied as they interacted with the prototype in situ. This approach enabled close attention to navigation behaviour, interaction with interface elements, and moments of hesitation or confusion as they emerged during use.

Observation notes were subsequently condensed into analytic categories focusing on:

1. navigation and movement patterns,

2. interaction with menus, toggles, and spatially anchored links, and

3. instances of interpretive uncertainty, such as difficulties distinguishing between physical reconstructions and digital elements.

These categories were used to contextualise survey responses and to guide design revisions between prototype iterations.

5.1. Usability and interaction clarity

Across both rounds, participants generally understood the basic premise of walking through the landscape to activate content, but several interaction issues recurred:

• ◦ Gesture expectations vs. implemented controls. Participants frequently attempted pinch-to-zoom and other common touchscreen gestures, indicating a mismatch between expected interaction conventions and the prototype’s button-based controls. This led to repeated trial-and-error and occasional requests for guidance.

• ◦ Visibility constraints in outdoor conditions. Bright sunlight during the September round reduced screen legibility and made subtle visual differences difficult to perceive. Conversely, the January round introduced physical constraints (snow, cold) that affected handling and movement (see Figures 6 and 7).

• ◦ Ambiguity in architectural toggles. The prototype included a toggle for alternative roof reconstructions. While exterior differences were visible, interior changes were often difficult to detect due to limited contrast and lighting. This created uncertainty about whether the toggle had produced meaningful change, particularly under glare.

Figure 6

Figure 7

These issues were reflected in observation and participant comments, and they informed revisions between test rounds, including clearer prompts and adjustments to presentation.

5.2. Object–place comprehension: what improved and why

The prototype’s central goal was to help users connect indoor exhibition artefacts to outdoor archaeological contexts. In both rounds, participants interacted with selected artefacts presented as rotaTable 3D models (e.g., an oval brooch, a glass goblet, and a gold foil figure) placed within reconstructed environments intended to reflect plausible contexts of use.

In the first test round, several participants expressed uncertainty about how the digital objects related to the physical exhibition. Observation suggested that spatial placement and 3D inspectability alone did not reliably communicate the intended connection. Users could engage with objects as “interesting digital items” without necessarily mapping them to the museum’s displayed artefacts.

Between the two test rounds, targeted audio cues were introduced to explicitly connect the digital models to their indoor exhibition counterparts and to clarify why the objects appeared in their respective locations.

In the second test round, after these changes, participants more frequently reported that the experience clarified the relationship between exhibition objects and their original locations. In other words, explicit narrative linking improved object–place comprehension more than additional detail or spatial realism alone.

Table 3 presents responses to two survey items used in the first and second test rounds. In the first test round, the item assessed the perceived importance of understanding the relationship between exhibition objects and their archaeological site. In the second test round, the item assessed participants’ understanding that the digital objects encountered on site corresponded to artefacts displayed in the museum exhibition.

5.3. Hybridity as a source of confusion: physical vs. digital structures

A persistent challenge involved the physical reconstructed Viking chieftain’s house, which stands adjacent to the original footprint. In both rounds, participants encountered a situation where the physical structure and a digitally reconstructed structure existed in close proximity.

• Round 1 observation: Participants frequently paused and looked for guidance when encountering the physical reconstructed house alongside the digital overlay. Instructions to “ignore the physical building” were not effective; the physical structure functioned as a dominant interpretive anchor that the digital layer had to compete with.

• Round 2 observation: Audio prompts helped, but confusion remained. Some feedback suggested that the experience would benefit from a clearly marked orientation element that acknowledges the physical reconstruction rather than treating it as noise.

These observations are summarised in Table 4.

5.4. Spatial misalignment and interpretive disruption

Minor misalignment occurred due to GPS drift of approximately one metre in some areas (see Figure 8). While users often tolerated small discrepancies in open terrain, misalignment became more disruptive where space was constrained or where digital elements appeared to “compete” with physical structures. In the interior context, for example, misaligned posts could appear to obstruct movement or draw attention away from interpretive content.

Figure 8

This suggests that hybrid museum settings impose different interpretive expectations than entertainment-oriented AR environments. In museums, both physical and digital reconstructions are commonly perceived as research-based and authoritative representations of the past, rather than as speculative or playful visualisations (Witcomb 2003: 102–120). As a result, visitors tend to interpret spatial accuracy as a marker of scholarly credibility. When a digital reconstruction appears misaligned with visible physical structures or with the surrounding landscape, even by a small margin, this can be read not merely as a technical limitation but as an interpretive inconsistency.

5.5. Summary of user feedback across rounds

Written feedback and observational patterns are synthesised in Table 5, which highlights recurring themes (imaging quality, object representation, narrative clarity, and the need for explicit linking between indoor exhibition and outdoor context).

6.1. Answering RQ1: Establishing a technical and interpretive link

RQ1 asked how a link can be established between archaeological remains at Borg and museum objects on display. From a technical perspective, the prototype connected these domains through sensor-driven positioning, anchored reconstructions, and spatial interaction points that linked outdoor locations to interpretive content and interactive artefacts. However, the evaluation demonstrates that technical linkage alone does not guarantee interpretive linkage. Establishing a meaningful connection between objects and place requires explicit cues that help users identify digital objects as corresponding to physical artefacts and understand why they appear in specific locations.

In practice, the most effective mechanism for establishing this connection was not increased visual realism, but clear narrative articulation. In the first test round, several participants struggled to link digital objects to their indoor exhibition counterparts, despite interacting with historically informed reconstructions. This gap was reduced in the second test round through the introduction of targeted audio cues that explicitly articulated the relationship between the digital models and the physical exhibition objects. These findings indicate that spatial placement alone is insufficient without narrative reinforcement, and that mobile AR can establish a meaningful object–place connection only when spatial anchoring is paired with concise, explicit explanation.

6.2. Answering RQ2: Enhancing understanding and engagement

RQ2 asked how such a link can enhance users’ understanding and engagement. The findings suggest three conditions under which object–place comprehension improves in hybrid museum environments.

First, narrative scaffolding at the moment of encounter proved critical. Participants benefited when the system clearly stated that the digital objects encountered on site corresponded to artefacts displayed in the indoor exhibition and explained the relevance of their spatial placement. This supports previous work highlighting the importance of narrative cues in situated and hybrid AR simulation (e.g. Sierra et al. 2024).

Second, visual differentiation in hybrid contexts emerged as a key factor. When physical reconstructions and digital overlays were similar in form or located in close proximity, participants struggled to interpret which elements belonged to which temporal layer. Without clear differentiation, such as translucency, labelling, or staged visual reveal, the experience risked being perceived as contradictory rather than complementary. This finding aligns with prior research showing that misalignment and low visual legibility can undermine trust and meaning-making in heritage AR (Bekele et al. 2018; Boboc et al. 2022).

Third, stable spatial anchoring where credibility is at stake was essential. Even minor positional inaccuracies undermined interpretive confidence when participants interpreted reconstructions as evidential representations of the past. In such contexts, positional precision functions not merely as a usability concern but as an interpretive requirement. Notably, these findings contrast with Wither, Tsai, and Azuma’s (2011) observation that users are generally tolerant of spatial mismatch in mobile AR. In the present study, even small misalignments produced disproportionate interpretive disruption when digital reconstructions appeared in close proximity to physical structures. This suggests that tolerance for spatial mismatch is lower in museum contexts, where reconstructions are often perceived as authoritative rather than speculative, and where visitors expect a high degree of spatial and historical credibility.

6.3. Designing for hybridity: Acknowledging, not hiding, the physical reconstruction

A particularly persistent source of confusion concerned the physical reconstruction of the Viking chieftain’s house, which stands adjacent to the original archaeological footprint. In the first test round, participants were instructed to disregard the physical structure, an approach that proved ineffective. Observations showed that participants repeatedly paused, hesitated, or sought clarification when encountering the physical building alongside its digital counterpart. This indicates that ignoring prominent physical features in hybrid environments is not a viable interpretive strategy.

Instead, the findings suggest that hybridity should be treated as a first-order design condition. The physical, reconstructed house functions as a strong interpretive signal and must be acknowledged rather than suppressed. A more effective strategy is to integrate the physical reconstruction into the digital narrative and provide explicit orientation cues that clarify representational and temporal layers.

One design implication emerging from participant feedback is the introduction of a clearly marked orientation element, such as a semi-transparent or “ghosted” representation of the chieftain’s house at the beginning of the simulation. Activated via GPS, this representation could function as an orientation aid that explicitly communicates the relationship between the physical reconstruction, the original archaeological footprint, and the digital model, gradually fading as users move away from the physical structure. Such an approach acknowledges the presence of the reconstruction while clarifying its interpretive status within the simulation. A preliminary concept illustrating this strategy is shown in Figure 9.

Figure 9

6.4. Spatial misalignment and technical implications

Spatial misalignment also affected user experience, particularly in the interior of the reconstructed house, where virtual load-bearing posts occasionally obstructed movement due to GPS drift of approximately one metre. While participants generally tolerated these discrepancies in open outdoor areas, they became problematic in confined spaces and contributed to interpretive confusion. This finding reinforces the observation that small positional errors can have disproportionate interpretive effects in hybrid museum environments.

To address this limitation, future iterations of the prototype will incorporate fiducial markers and ARKit-based tracking to improve spatial precision. Importantly, this represents a targeted response to a specific interpretive breakdown identified during testing, rather than a generic technological upgrade.

6.5. Limitations

This study is based on formative evaluation with a modest sample (N = 31) recruited through convenience sampling. The aim was not statistical generalisation, but to identify recurring patterns of use, interpretive breakdowns, and design challenges during real-world deployment of a mobile AR prototype. Accordingly, the analysis focuses on descriptive trends across test rounds rather than inferential statistics.

The two test rounds were conducted under substantially different outdoor conditions, including seasonal variation, temperature, snow cover, and lighting. These factors affected usability, for example, screen visibility and movement in the terrain, and may also have shaped participants’ interpretive responses. While this variability complicates direct comparison, it also reflects the realities of deploying mobile AR in outdoor heritage settings and provides valuable insight into how environmental conditions interact with design choices, such as the impact of strong sunlight and screen reflections illustrated in Figure 10.

Figure 10