Introduction
The Helium blockchain uses a novel work algorithm called "Proof-of-Coverage" ("PoC") to verify thatHotspots are located where they claim. Put another way, PoC tries to verify, on an ongoing basis,that Hotspots are honestly representing their location and the wireless network coverage they arecreating from that location.
While current 5G Hotspots are also able to participate in the IoT Network and PoC, the CBRS SmallCell and 5G aspects of a 5G Hotspot require a new approach to validating PoC.
Modeled Coverage introduces a fundamental difference in using hexes for the Mobile Network comparedto the IoT Network where hexes determine the locations and number of Hotspots within the given hex.In the Mobile Network, hexes determine where the coverage exists. Modeled Coverage data allows us toun-tie a Radio's location from its signal strength and propagation.
One Way Traffic
A CBRS Small Cell Radio ("Radio") can only operate as a transmitter; they cannot receive incomingdata and thus can not "hear" one another directly.
LoRaWAN Hotspots utilize a PoC method that benefits from Hotspots being used as senders ("Beacon")and receivers ("Witness") of PoC activity; read more aboutProof-of-Coverage
Proof of Coverage Milestones
Determining useful coverage and the impact on MOBILE PoC rewards will be based on a number offactors:
These factors will be rolled out in stages over the next twelve months and explained in detail toensure Network Builders can take concrete steps to plan or improve their setups. Check out theMOBILE PoC Roadmap.
Reward Algorithm
The proposed new algorithm for MOBILE Reward calculation in the MOBILE Rewards Oracle is as follows:
- Supply the declared transmitter power of each Radio and its location to the Obstruction DataOracle.
- Get all hexes that have coverage from Outdoor Radios based on the information from theObstruction Data Oracle.
- Based on the location of Indoor Radios, get all hexes with Indoor coverage and all adjacenthexes.
- Use projected signal loss information from the Obstruction Data Oracle to determine the potentialsignal strength of each Outdoor Radio in each hex.
- For each hex, get at most 5 Outdoor Radios with the top signal strength of the same level.
- If there are more than 5 Radios with the same signal strength level, use the
coverage_claim_time
value to determine the top 5 oldest installations wherecoverage_claim_time
is the timestampwhen the Radio received the spectrum access grant for the first time. - To prevent rewarding "dead" Radios, we propose to reset
coverage_claim_time
if the Radio was notgenerating a Heartbeat for more than 72 hours and use the time of the last Heartbeat as the newcoverage_claim_time
.
- Get max 5 Indoor Radios using the same approach as above for Outdoor Radios.
- Based on Modeled Coverage Reward Tiers, sum up all estimatedcoverage points earned by each Radio in all hexes and multiply that by
speedtest_multiplier
foreach Radio. - Divide the total number of MOBILE Rewards emitted during the Rewards Period by the sum ofmultiples of heartbeat_multiplier,estimated_reward_points, andspeedtest_multiplier for each Radio to determine MOBILE Rewards per
estimated_reward_point
for Radio with a passing Speed Test and valid Heartbeats. - Multiply the MOBILE per
estimated_reward_point
by the sum of each Radio's reward points todetermine the MOBILE Rewards for each Radio.
The new formula for Reward calculation per Radio:
$$W = k_H \times k_S \times C_E \times W_p$$
Variable | Description |
---|---|
$W$ | Total rewards |
$k_H$ | Heartbeat multiplier |
$k_S$ | Speed test multiplier |
$C_E$ | Estimated coverage points |
$W_p$ | Epoch rewards per coverage point |
Calculation Example
For simplicity, assume that the total MOBILE Rewards per period is 10,000.
Radio | Heartbeat | Heartbeat multiplier $k_H$ | Speed Test | Speed test Multiplier $k_S$ | Hex 1 - Hex 10 | Hex 11 - Hex 20 | Hex 21 - Hex 220 | Total Coverage Points | Total Reward Points |
---|---|---|---|---|---|---|---|---|---|
1 (Outdoor) | Ok | 1 | Acceptable | 1 | 160 | 80 | 800 | 1,040 | 1,040 |
2 (Outdoor) | Ok | 1 | Poor | 0.25 | 80 | 40 | 120 | 30 | |
Points | 1,070 |
Table 3. Simplified data for two Outdoor Radios with Heartbeat, Speed Test, and EstimatedCoverage Points for one Reward Period.
Radio | Heartbeat | Heartbeat multiplier $k_H$ | Speed Test | Speed test Multiplier $k_S$ | Hex 1 | Hex 2 | Hex 3 | Hex 4 | Total Coverage Points | Total Reward Points |
---|---|---|---|---|---|---|---|---|---|---|
3 (Indoor) | Ok | 1 | Degraded | 0.5 | 100 | 100 | 100 | 400 | 700 | 350 |
Points | 350 |
Table 4. Simplified data for one Indoor Radio with Heartbeat, Speed Test, and Estimated CoveragePoints for one Reward Period.
$$R \times (1 \times 1 \times 1040 + 1 \times 0.25 \times 120 + 1 \times 0.5 \times 700) = 10,000$$
Where R is the reward per one estimated coverage point.
$$R = 6.54$$
Calculation | Total MOBILE Rewards | |
---|---|---|
Radio 1 (Outdoor) | 6.54 x 1 x 1 x 1040 | 6,806.30 |
Radio 2 (Outdoor) | 6.54 x 1 x 0.25 x 120 | 903.14 |
Radio 3 (Indoor) | 6.54 x 1 x 0.5 x 700 | 2,290.58 |
Modeled Coverage
With thepassing ofHIP-74Modeled Coverage aims to allow for more appropriate Rewards and incentivizes the deployment ofRadios at optimal locations.
This new algorithm uses the location of the Radio to calculate MOBILE Rewards based on the number ofres12 hexes receiving coverage from a given Radio.
The MOBILE Rewards Oracle will calculate a coverage model for each Radio by considering the physicallocation, antenna direction, radio max transmit power, and Obstruction Data Oracle inputs tocalculate the number of res12 hexes receiving coverage from each Radio.
Depending on the number of res12 hexes and the resulting signal power in each, the MOBILE RewardsOracle will allot Coverage Points to a given Radio for calculating its MOBILE Rewards instead ofassuming a fixed Radio Type multiplier Reward.
Reward Hexagon Sizing
Analysis of the coverage by a single Outdoor Radio with Modeled Coverage data shows that the res8hex used in the IoT Network is too large to represent a Radio's coverage. Using res12 hexes betteraligns with what an Outdoor Radio can reasonably cover, with an average area of 0.0003071 km² and anedge length of 9.4 m, slightly bigger than an average single-family home.
To summarize:
- Hexes in the Mobile Network identify the level of coverage from all the different Radios providingcoverage to each res12 hex.
- A Radio can provide coverage for a hex it is not located within.
- A Radio can earn Rewards by providing coverage in multiple hexes.
To help visualize the significant difference between these two resolutions, below is a map with agreen res8 hex and a purple res12 hex inside.
Modeled Coverage Reward Tiers
info
Each Radio is given a coverage points value depending on the number and signal strength of theres12 hexes receiving coverage from the Radio.
Outdoor Radios
Modeled Coverage replaces Radio Type multipliers with coverage points represented by four tiersof potential signal power as modeled by the Obstruction Data Oracle - High
, Medium
, Low
, andNone
.
The reward tiers for Outdoor Radios will be as follows:
Tier 1 | Tier 2 | Tier 3 | Tier 4 | |
---|---|---|---|---|
Potential Signal Power | $P > -95 dBm$ | $-95 dBm \ge P > -105 dBm$ | $-105 dBm \ge P > -115 dBm$ | $P \le -115 dBm$ |
Potential Signal Level | High | Medium | Low | None |
Estimated coverage points | 16 | 8 | 4 | 0 |
Modeled Coverage Distance Limit for Outdoor Radios
The Modeled Coverage algorithm limits the radius of coverage for a given Radio to 1000 meters. Thisvalue results from evaluating a selection of "ideal case" Radios, which provided usable coverage toroughly 500 meters and doubled that value. A larger radius would also be too heavy for thecomputation and is a current trade-off.
Indoor Radios
Like Outdoor, Indoor Radios can collect MOBILE Rewards by providing coverage in multiple hexes.
The Obstruction Data Oracle cannot evaluate coverage by Indoor Radios as there are no good andreliable data sources about obstacles in indoor spaces. Instead, the Modeled Coverage Oracle uses anapproximation based on the data gathered while testing certified Indoor Radios in various indoorsettings.
From this testing, the hex in which an Indoor Radio is physically located receives maximum signalstrength coverage, and all adjacent hexes receive lower signal strength. Therefore, the algorithmerrs on generosity to ensure equitable potential MOBILE Rewards for Indoor Radio coverage.
Estimated per-hex coverage points for Indoor Radios are intentionally significantly higher thanthose awarded to Outdoor Radios to balance the importance of Indoor Radios in the Mobile Networkwith their relatively smaller coverage footprints.
The reward tiers for Indoor Radios will be as follows:
Tier 1 | Tier 2 | |
---|---|---|
Location | Inside hex | All adjacent hexes |
Potential Signal Level | High | Low |
Estimated coverage points | 400 | 100 |
These values were chosen by taking all the Outdoor Radios in an example area and analyzing theresulting distribution of estimated coverage points per Radio. Estimated coverage point values werechosen such that Indoor Radios receive roughly ¼th the estimated coverage points as the 95thpercentile Outdoor Radios and about ½ the points of the average Outdoor Radio.
These results align with the current Radio Type Multipliers approach 1:2.5:4
for Indoor Radios,Outdoor Radios, and High-Power Outdoor Radios. Figure 2 below is the cumulative distributionfunction plot of the sample market Outdoor Radio estimated coverage points.
Cumulative distribution function of Outdoor Radio estimated coverage points for 904 Radios in theL.A. market area.
Modeled Coverage Explorer
check coverage
View mobile Hotspot coverage on the Modeled Coverage Explorer.
Modeled Coverage data will play a key role in providing information to visualize it, with more dataadded as additional external sources (e.g., Mobile Mappers) come intoplay.
HIP-74 proposed the creation of a new Mobile Explorer dedicated exclusively to the Mobile Network asadding more features and data for MOBILE Rewards to the IoT Network Explorer becomes practicallyimpossible as the Mobile Network grows. The first iteration of the Mobile Explorer proposes a basicmap overlay of 5G data coverage with signal strength.
Uptime Heartbeats
A "heartbeat" is data sent by the 5G Hotspot indicating that a connected Radio is authorized totransmit 5G coverage.
Heartbeats occur every 60 seconds and transmit authorizations are valid for 240 seconds. Thisrolling overlap ensures that a Small Cell Radio can broadcast on an ongoing basis while allowing forbrief interruptions.
To encourage reliable signal availability, 5G Hotspots and connected CBRS Small Cell Radios mustgenerate at least one valid heartbeat per hour in 12 unique hours of the 24-hour Reward Period.
Heartbeat Reward Tiers
Heartbeat impact on Radio rewards
Each Radio is given a heartbeat multiplier of either 0
or 1
.
Each Radio is awarded 1 point for a valid heartbeat in each hour, with a maximum of 24 points in theReward Period. All Radios with at least 12 points in the 24-hour Reward Period are given aheartbeat multiplier
of 1
and will be eligible to earn MOBILE Rewards.
Heartbeat Examples
Choose from the options below to see some heartbeat example scenarios and the impact on MOBILErewards.
- ✅ No Missed Heartbeats
- ✅ Half Day Offline
- ⛔️ Mostly Offline
The Radio generated a heartbeat every hour of the day, earning 24 points.
This Radio will be eligible for rewards during the Reward Period.
Backhaul Speed Tests
A 5G Hotspot automatically performs Speed Tests, with the results averaged by the MOBILE PoC Oracle.The MOBILE PoC Oracle includes these Speed Test metrics in the Rewards calculations for each 5GHotspot.
The 5G Hotspot will randomly measure download, upload, and latency to the MOBILE verifier twice in agiven 24-hour period. If a test can not complete successfully, the 5G Hotspot automatically retriesevery 30 minutes until a Speed Test is successful. The first Speed Test for a newly installed 5GHotspot is performed within the first 12 hours after powering it on.
A minimum of two and maximum of six Speed Tests are required in the previous 48 hours to calculate aMoving average results for each download, upload, and latency, after which a new series of SpeedTest results must be collected before a Moving Average can be calculated.
Speed Test Reward Tiers
Speed test impact on Radio rewards
Each 5G Hotspot is given a speedtest multiplier
of either 0.0
, 0.25
, 0.5
or 1.0
.
This multiplier is applied to all connected Radios in rewards calculations.
Many locations where CBRS connectivity is being deployed, including some rural areas, do not alwayshave the high-speed Internet connectivity needed to meet the acceptable Internet requirements forGenesis rewards consistently.
Often these areas do not have good cellular coverage either. That's why we believe it is essentialto incentivize Helium deployments in less well-connected areas.
In response to community feedback and further exploration by Chris from DeWiPulse for hisanalysis of and commentary on public Internet data in the United States,Speed Test results are categorized into one of four Tiers - Acceptable
, Degraded
, Poor
, andFail
.
Speed Test Tier | Speedtest Multiplier | Requirement (speeds in Mbps, latency in ms) |
---|---|---|
Acceptable | 1.00x | 100+ Download, AND 10+ Upload, AND <50 Latency |
Degraded | 0.50x | 50+ Download, AND 5+ Upload, AND <75 Latency |
Poor | 0.25x | 30+ Download, AND 2+ Upload, AND <100 Latency |
Fail | 0.00x | <30 Download, OR <2 Upload, OR >100 Latency |
Tiered Speed Test values are used as a multiplier in Rewards calculations as follows:
- Speed Test Results are put into Tiers based on the minimum value of each Download, Upload, andLatency (logical
AND
). - Speed Test results that do not meet the minimum requirements for any Download, Upload, or Latencyare considered to have
Failed
and are not eligible for Rewards until the Speed Test Average isimproved (logicalOR
).
Requirements Will Evolve
As the needs and realities of the Network change, backhaul requirements will evolve to better suitthe Network and its users.
Speed Test Examples
Choose from the options below to see some speed test example scenarios and the impact on MOBILErewards.
- ✅ All Acceptable
- ❎ Degraded Upload
- ⛔️ Poor Download
- 🆘 Failed Latency
Download 200 -> `Acceptable`
Upload 40 -> `Acceptable`
Latency 15 -> `Acceptable`
---------------------------------
Speed Test Tier -> `Acceptable` -> 1.0x Multiplier
Speed Test Example Calculations
Here are a set of tables to help visualize the components that make up the Speed Test result.
info
It may take up to 12 hours for newly connected 5G Hotspots to perform a Speed Test, and a minimum of2 Speed Tests are required for the MOBILE PoC Oracle to calculate a moving average.
- ✅ Successful Speed Test
- ✅ Failing A Speed Test
- ❎ Missing A Single Speed Test
- ❎ Multiple Sub-Optimal Speed Tests
- 🆘 Missing Multiple Speed Tests
This 5G Hotspot is passing all Speed Test metrics and is providing a high-quality connection.
Test 1
This 5G Hotspot will not be eligible for rewards during the Reward Periods for which the resultsof Test 1 are the most recent as there is not enough data to calculate a Moving Average.
Test 2 and Later
This 5G Hotspot will be eligible for full rewards during the Reward Periods for which theresults of these are the most recent.
Statistics on Helium Explorer
Speed Test metrics are viewable on the 5G Hotspot's page on Explorer under the 5G Statistics
tab.Here are examples of each status and a brief description of each metric and the resulting impact onRewards calculations.
Helium Explorer Examples
Choose from the options below to see some example scenarios and the impact on MOBILE rewards.
- ✅ Speed Test Passing
- ❎ Degraded Latency
- ❎ Poor Download
Speed Test Passing - Acceptable
The Moving Average for all Speed Test metrics is in the Acceptable
range.
A 1.0x
multiplier is applied for the current Reward Period.
- 🆘 Speed Test Fail
- ⏺️ Not Measured
- ⏺️ No Recent Data
- ⏺️ Not Enough Data
Speed Test Fail - Download and Upload
The Moving Average for Latency metrics is in the Acceptable
range, however the 10 Mbps Download and 1 Mbps Upload metrics are below the Poor
Tier.
A 0.0x
multiplier is applied for the current Reward Period.