BLE vs LoRaWAN vs LTE-M Trackers for Pallets and Containers

The pilot looked perfect on the dashboard: 400 returnable containers, a dot for each one, refreshed every few minutes. Six weeks into live use, a third of the dots had not moved for nine days. Some trackers had flat batteries, some sat inside steel stacks that blocked the signal, and some had been moved to other containers by depot staff without anyone updating which tracker belonged to which asset.

The radio choice decides most of that outcome before a single device ships. Bluetooth Low Energy (BLE), LoRaWAN and LTE-M differ in where a tracker can be heard, how many bytes it can send, how long a battery lasts and what each position costs. Matching the radio to where the asset actually spends its days turns a tracking pilot into a ledger that operations can trust.

• What this guide covers: coverage models, US and Indian radio rules that cap airtime and power, payload budgets, battery arithmetic, cost per reading, pairing records and how to tell a dead tracker from a coverage gap.
• What it adds: worked airtime and battery formulas, a three-year cost comparison in dollars and rupees, a pairing failure traced hour by hour and a pilot checklist to run before a fleet purchase.

Three radios, three answers to where an asset can be heard

Pick the radio from the asset's itinerary, not from the device datasheet. BLE is heard only near a reader you own; LoRaWAN is heard near a gateway, yours or a network operator's; LTE-M is heard wherever a mobile operator has deployed it. Everything else in the comparison follows from that.

• BLE tags: small, cheap beacons that advertise their identity to fixed gateways, handheld scanners or phones. They report location only as "near reader X", so a pallet is visible at your docks and invisible on the road unless the vehicle carries a gateway.
• LoRaWAN trackers: long-range, low-rate radios that send short uplinks to gateways several kilometres away in open terrain. Coverage is what you or a network operator build, which suits yards, ports, plants and fixed corridors.
• LTE-M trackers: cellular devices on operator networks under 3GPP Release 13 and later, with power-saving modes. They work across cities and highways where the operator has enabled LTE-M, and they buffer readings when they cannot attach.

Most real fleets mix them. A common pattern puts BLE tags on pallets and a cellular gateway on the trailer, so the trailer reports where it is and which tagged pallets are inside. Another puts LoRaWAN on yard equipment and containers within a port, where one gateway on a light mast covers the whole stack area.

Our guide to a returnable pallet, crate and cage ledger explains how to balance custody of those assets between sites. This article answers a narrower engineering question that comes before it: which radio will deliver enough position readings, at an acceptable cost, to feed that ledger without manual scans.

Radio rules that cap airtime and power in the US and India

LoRaWAN and BLE run in licence-exempt bands, so national rules decide how long a tracker may transmit and at what power. LTE-M runs in operator-licensed spectrum, where the device needs certification but airtime is managed by the network. The two sub-gigahertz rule sets differ enough to change a tracker's firmware between the US and India.

In the US, 47 CFR 15.247 governs the 902-928 MHz band. For frequency-hopping systems with hopping channels narrower than 250 kHz, it requires at least 50 hopping frequencies and limits average occupancy on any frequency to 0.4 seconds within a 20-second period, with peak conducted output up to 1 watt. The LoRa Alliance's regional parameters translate this into a 400 ms uplink dwell limit for US902-928 devices.

India's Use of Low Power Equipment in the Frequency Band 865-868 MHz for Short Range Devices (Exemption from Licence) Rules, 2021 (G.S.R. 853(E), 10 December 2021) list tracking, tracing and data acquisition devices in Table II at up to 500 mW e.r.p. with adaptive power control, a 200 kHz bandwidth and duty cycle limits of 10 percent for network access points and 2.5 percent otherwise.

LoRaWAN tracker firmware for India must respect Table II of G.S.R. 853(E), 2021: 500 mW e.r.p. with adaptive power control and a 2.5 percent duty cycle for end devices, while US902-928 firmware must keep each uplink under the 400 ms dwell limit derived from 47 CFR 15.247(a)(1)(i).

Fitting a useful position report into 11, 51 or 31 bytes

Design the payload before choosing the radio, because the slowest LoRaWAN data rate in the US carries only 11 bytes of application data. A position, a battery reading and a motion flag fit if they are packed as integers; a JSON message does not fit at all.

• Latitude and longitude: two signed 32-bit integers in units of 1e-7 degrees take 8 bytes and keep roughly centimetre resolution, far finer than a GNSS fix on a container stack needs.
• Battery voltage: one byte holding millivolts above 2,000 in 10 mV steps covers 2.00 to 4.55 V, enough to trend a lithium primary cell.
• Status flags: one byte holds motion, shock, door, GNSS-valid and buffered-reading bits, leaving three spare.
• Fix age: one byte in minutes up to 255 tells the platform how stale the position was when sent, which matters for buffered cellular readings.

Those fields total 11 bytes, exactly the US DR0 limit with no room for a sequence counter, which LoRaWAN already provides in its frame header. India's IN865-867 profile allows 51 bytes at its slowest rate, so the same device can add temperature and a second motion window there. BLE legacy advertising carries up to 31 bytes, but most tag designs send only an identifier and let the gateway add the location.

Battery arithmetic: sleep current, airtime and reporting interval

Battery life is set by three numbers: sleep current, energy per report and reports per day. Datasheet claims of "up to five years" assume one report a day; a tracker sending hourly with a GNSS fix will last months, not years. Run the formula with your own interval before comparing devices.

LTE-M devices save energy with Power Saving Mode and extended discontinuous reception. The GSMA's LTE-M deployment guide (CLP.29) recommends networks accept a T3412 extended timer from 4 hours up to the 3GPP Release 13 maximum of 413 days, a T3324 active timer of at least 16 seconds, and eDRX cycles from 5.12 seconds to 43.69 minutes. A tracker that only sends never needs to listen between reports.

1) LoRa uplink airtime (Semtech time-on-air formula, explicit header, CRC on, CR 4/5)
   Tsym      = 2^SF / BW
   n_payload = 8 + max( ceil( (8*PL - 4*SF + 28 + 16) / (4*(SF - 2*DE)) ) * 5 , 0 )
   ToA       = (n_preamble + 4.25 + n_payload) * Tsym
   PL = application bytes + 13 (MAC header, frame header, port, MIC)

   US DR0: SF10, BW 125 kHz, DE 0, PL = 11 + 13 = 24, preamble 8
     Tsym = 8.192 ms; n_payload = 8 + 5*5 = 33; ToA = 45.25 * 8.192 = 370.7 ms  (< 400 ms dwell)
   IN DR0: SF12, BW 125 kHz, DE 1, PL = 20 + 13 = 33, preamble 8
     Tsym = 32.768 ms; n_payload = 8 + 7*5 = 43; ToA = 55.25 * 32.768 = 1,810 ms
     2.5% duty cycle over 1 hour = 90 s, so at most 49 such uplinks per hour

2) Battery life (illustrative LTE-M tracker, hourly report with GNSS fix)
   E_day  = I_sleep * 24 + n * (Q_tx + Q_fix)            [mAh per day]
   L_days = C * k / E_day
   I_sleep = 0.005 mA (PSM), Q_tx = 0.25 mAh, Q_fix = 0.10 mAh, n = 24
   E_day  = 0.12 + 24 * 0.35 = 8.52 mAh
   C = 3,600 mAh primary cell, k = 0.80 derating for cold and ageing
   L_days = 2,880 / 8.52 = 338 days

The airtime result explains why US LoRaWAN trackers fall back to tiny payloads at long range: a 24-byte frame at SF10 already uses 371 of the 400 permitted milliseconds. The Indian calculation shows that the 2.5 percent duty cycle is rarely binding for a tracker reporting every 15 or 30 minutes, but it does bind a device that retries aggressively after missed acknowledgements.

Cost per reading across hardware, connectivity, gateways and swaps

Compare radios on cost per delivered reading over the tracker's life, not on device price. A cheap tag that is heard only at your own docks has a tiny cost per reading and no readings at all in transit, so the useful denominator is readings from the places where decisions are made.

The table below models 2,000 returnable assets over 36 months. Device, gateway, data plan and labour figures are illustrative assumptions, not vendor quotes; replace them with your own tenders. Delivered readings count only reports that reach the platform: BLE when an asset is at one of six instrumented sites, LoRaWAN with 70 percent delivery across mixed coverage, LTE-M with 97 percent after buffering.

Tracker cost per delivered reading for 2,000 returnable assets over 36 months ranges, in an illustrative model, from $0.00025 for BLE tags heard only at six sites to $0.0067 for hourly LTE-M trackers, but only the LTE-M option reports during linehaul, where detention and loss decisions are actually made.

The India LoRaWAN line assumes a private network because public coverage varies by city; the US line assumes a public network subscription. Both are design choices you should test, not facts about either market. Notice that the LTE-M battery line is as large as the device price once three replacements are counted, which is why the reporting interval is the biggest cost lever.

Pairing records: which tracker is on which pallet or container

A tracker reports its own identity, never the asset's. The pairing between a device ID and a container number or returnable asset identifier is a business record with an owner, a start time and an end time, and it causes more tracking errors than any radio limitation.

The pairing record needs a paired-to timestamp because trackers are moved. When a container returns empty and its tracker is fitted to the next outbound unit, the platform must close the old pairing at that moment. Otherwise every later position is credited to the wrong asset, and a demurrage or loss decision is made on the wrong evidence.

1. 08:10, empty return: container NXLU 482193 arrives at a Nagpur depot with tracker DEV-7C41; gate-in is recorded in the yard system as event GI-55021.
2. 09:45, refit: a depot hand removes DEV-7C41 and mounts it on outbound container NXLU 510774 but scans nothing, so pairing PR-3318 still links DEV-7C41 to NXLU 482193.
3. Day 2, 06:30, false movement: the platform shows NXLU 482193 travelling towards Raipur; the customer service desk tells the lessor the empty unit is on hire.
4. Day 6, 15:20, invoice: per-diem billing run PD-0921 charges six days of hire on NXLU 482193, which never left the depot, while NXLU 510774 shows no movement at all.
5. Controlled version: a rule that a tracker reporting motion while its paired asset has a gate-in event and no gate-out raises a pairing exception within 30 minutes would have flagged DEV-7C41 at 10:15 on day 1.

The fix is a scan at refit, closing one pairing and opening another in a single transaction, plus a cross-check against custody events. The radio was never the problem in this trace; the readings were accurate and simply attached to the wrong container.

Silent trackers: a flat battery, a steel stack or a coverage hole

Treat silence as its own state, with a rule for when it starts and evidence for why. A missed report can mean a dead battery, a tracker shielded by steel, an area without coverage or a device removed from its asset, and each needs a different response.

Our article on separating a temperature alarm from a sensor-data failure sets out the same problem for reefer probes. For position trackers the evidence is different: last battery voltage, last signal quality, whether the last report showed motion, and whether other trackers at the same site are still reporting.

Models are useful where the evidence is probabilistic, such as signal strength indoors. They should never move a custody record on their own, because a demurrage, hire or loss charge follows from that record. Keep the fixed rules simple enough that a depot supervisor can explain them to a customer in one sentence.

Matching the radio to yards, linehaul lanes and ocean legs

Choose per lane and per asset type, then accept a mixed fleet. A single radio across every asset usually overpays in the yard or goes blind on the road.

• Closed-loop pallets between your own sites: BLE tags with gateways at each dock, because readings are needed only at receipt and despatch, and the custody ledger, not live position, settles balances.
• Port and plant yards: LoRaWAN, because a few masts cover a large fixed area, payloads are small and nothing leaves the site often.
• Domestic road linehaul and rental containers: LTE-M, because position on the road drives detention, theft and per-diem decisions, provided your operator has LTE-M on the lanes you run.
• Ocean legs: any cellular tracker loses the network at sea and buffers readings until it reaches port. If you need sea positions, use carrier events from a tracking API instead, as covered in our guide to container tracking API events.
• Hybrid trailers: BLE tags on pallets plus one cellular gateway in the trailer cut the per-asset cost while still reporting position in transit.

LTE-M trackers for rental containers should request a T3412 extended timer and eDRX cycle within the GSMA CLP.29 recommended ranges, 4 hours to 413 days and 5.12 seconds to 43.69 minutes, and report on a 60-minute interval unless detention terms justify faster reporting and the extra battery swaps.

A pilot and evaluation checklist before a 10,000-tracker order

Run a 90-day pilot with at least 100 devices per candidate radio on your real lanes and assets before any volume order. Measure delivered readings, battery voltage trend and pairing exceptions, and score each option on evidence from the pilot rather than on the vendor's specification sheet.

Score the pilot on the questions your finance and operations teams actually ask, such as how many detention disputes were settled with tracker evidence and how many pallets were found that would otherwise have been written off. Those counts, not dashboard refresh rates, decide whether a fleet order pays back.

More articles on asset visibility and returnable packaging sit on the Logistics & Supply Chain hub. If you are writing a specification for a tracking platform that holds device registries, pairing records and custody events together, share your requirements with us and we will respond with a scoped proposal.