MenuClose
A Heimdall sensor standing on a forest path at Klåveröd, the path running on between tall spruces.
In field testing

The Field Log.

Heimdall is tested in the field, usually in forests. Whatever I put it through, I go over the results back home and publish them here, misses and false alarms included.

Updated 16 Nov 2026

Simulation 01 ·

A person walking.

3 × 171 walks Calm, wind, rain
Detected: 166–⁠170, of 171 walks per run. False alarms: , not measured. Typical alarm: 38–⁠42m, 123–⁠137 ft.

A simulated person walks up to the sensor from 70 m (230 ft). The footsteps are real ones from the field days, placed into 19 recordings of the ground with nobody around, in calm, wind and rain. Each gets nine walks, 171 in all, and the set was run three times with other footsteps drawn each time, so the results are ranges.

Field Day 06 ·

Jällabjär.

Grass, dirt, gravel Wind, rain
Detected: 4 of 5, walks. False alarms: 0, walks and 31 min of noise. Typical alarm: 49m, 31 s of warning, 160 ft.

5 walks on grass, dirt and gravel. A cold and windy afternoon with occasional rain. Wet ground with mostly grass, moss, mud, leaf litter, and some rocks. Both off-trail and on-trail.

No false alarms, during the walks or in 31 minutes of wind and background noise.

See Field Day 06
Field Day 05 ·

Kopparhatten.

Gravel, leaf litter Calm
Detected: 5 of 5, walks. False alarms: 0, walks and 25 min of noise. Typical alarm: 38m, 30 s of warning, 126 ft.

5 walks on gravel and soft leaf litter (off-trail). Sunny afternoon with almost no wind. No rain for the day. Rained the day before.

The ground was full of fallen leaves, very damp, mushy, and soft, with some moss under the leaves. Like walking on top of a mattress. The ground was mostly very uneven with many twigs, rotten logs, and with some elevation changes.

No false alarms, during the walks or in 25 minutes of wind and background noise.

See Field Day 05
Field Day 04 ·

Klåveröd.

Pine needles, leaf litter, gravel, grass Calm, wind
Detected: 8 of 8, walks. False alarms: 1, walks and 24 min of noise. Typical alarm: 35m, 23 s of warning, 114 ft.

8 walks on various trails with pine needles, leaf litter, gravel, and grass. Humid autumn day but no rain during the field session. Ground was damp and often covered with fallen leaves. Some strong wind gusts every now and then.

One false alarm during the walks, none in 24 minutes of wind and background noise.

See Field Day 04
Field Days 01–03 · 2025–2026

The first three.

Pine needles, gravel, leaf litter, grass Calm, rain
Detected: 30 of 31, walks. False alarms: 0, walks and 19 min of noise. Typical alarm: , not tracked.

The first three field days came before the walks were tracked by GPS, so they have no distances and no pages of their own. Their walks towards the sensor and their false alarms are counted in the totals above.

No false alarms, during the walks or in 19 minutes of wind and background noise.

Next

Nights, bears and more field days.

The first nights out and the first simulated bears, and more field days this winter. On the list: Traneröd’s Moor, Skrylle, Natthall and Lake Odin.

Follow along
Records

The best results so far.

The best results of all 18 walks tracked by GPS.

Farthest detection
98m322 ftKopparhatten, walk 01
Longest warning
47sKopparhatten, walk 01
Farthest alarm
78m256 ftKopparhatten, walk 01
Every step felt
69of 69 stepsKlåveröd, walks 01 and 02