4. Occupant classification and sensing for far-side deployment

Far-side airbags protect occupants whose bodies move toward the vehicle centerline during a lateral impact. Unlike a frontal airbag, which fires into a relatively predictable space between the occupant and the steering wheel or instrument panel, a far-side airbag must deploy into the gap between the driver and the front passenger, or between two rear-seat occupants. That gap changes constantly with belt usage, occupant size, and seating position. The restraint control module (RCM) therefore cannot rely on a single crash sensor alone; it must fuse crash severity with occupant-state information before deciding whether, when, and how aggressively to deploy the far-side device.

This section examines the three sensing inputs that dominate far-side deployment logic: seat belt status, occupant size classification, and occupant position detection. It explains what each sensor measures, how the signals are conditioned and interpreted, and how the RCM combines them into a deployment decision.

Seat Belt Status: The First Gating Condition

Seat belt status is the most direct indicator of whether an occupant is restrained and where their torso will travel during a side impact. A buckled belt changes the occupant's kinematics fundamentally: the torso is coupled to the seat and vehicle, so lateral excursion is reduced and the head-to-centerline distance is smaller. An unbuckled occupant, by contrast, can slide laterally across the seat and may already be displaced toward the centerline before the far-side airbag inflates.

Belt status is detected by a buckle switch, typically a Hall-effect or mechanical contact sensor integrated into the buckle stalk. The switch produces a discrete state: BUCKLED or UNBUCKLED. Some systems add a belt tension sensor (BTS) at the anchor to distinguish a lightly loaded belt from a properly tensioned one, and a belt payout sensor to estimate how much webbing has been extracted. The RCM debounces the raw switch signal and applies a plausibility check against vehicle speed and door state to reject transient states such as a belt buckled with no occupant present.

For far-side logic, belt status acts as a gating condition rather than a continuous variable. A typical rule set is:

• BUCKLED, occupant present: far-side deployment enabled with nominal threshold.
• UNBUCKLED, occupant present: deployment enabled but with a lower crash-severity threshold and, in some calibrations, a delayed or reduced-output deployment to account for the larger occupant excursion.
• No occupant detected: far-side deployment inhibited on that side.

Occupant Size Classification

Occupant size determines the mass and the effective radius of the body that the airbag must manage. A 5th-percentile female and a 95th-percentile male differ substantially in torso mass, shoulder height, and head position. The RCM uses this information to select an inflation profile, and in some architectures to suppress deployment when the occupant is a child in a rear-facing child restraint system (CRS).

Two sensing approaches dominate. The first is occupant weight sensing (OWS), which uses strain gauges or pressure-sensitive resistors in the seat pan and back frame. The system sums the measured load, subtracts seat structure tare, and classifies the result into bands such as EMPTY, CHILD, 5th female, 50th male, and 95th male. The second is capacitive or electric-field sensing, which measures the dielectric coupling between electrodes in the seat and the occupant's body. Capacitive systems are less sensitive to cargo load and can distinguish a human from a bag of groceries, but they require careful shielding against moisture and seat heating elements.

A representative classification table used during calibration is shown below. The values are illustrative teaching parameters, not measured production data.

Classification is not instantaneous. The RCM applies a moving average over a window of roughly 1–2 seconds and requires the class to remain stable before it is accepted. This prevents a momentary load shift, such as an occupant reaching for the glovebox, from reclassifying the seat.

Occupant Position Detection

Position sensing answers a different question: not how large the occupant is, but where the occupant's head and torso are at the moment of impact. Far-side deployment is especially sensitive to this because the airbag inflates into the centerline gap, and an occupant already leaning toward the centerline may be out of position (OOP) relative to the intended deployment zone.

Common position-sensing technologies include:

• Seat track position sensor: a Hall or potentiometer sensor that reports fore-aft seat location. A seat moved far forward reduces the distance to the centerline and changes the required inflation timing.
• Seat back angle sensor: reports recline angle, which affects torso and head trajectory.
• Ultrasonic or time-of-flight sensors: mounted in the seat back or headliner, they measure the distance to the occupant's head and chest, detecting lean-in, lean-out, and forward lean.
• Camera-based occupant monitoring: increasingly common in vehicles with driver monitoring systems; it provides head position, gaze, and posture at higher resolution than discrete sensors.

The RCM fuses these signals into a position estimate. A practical implementation defines a centerline proximity index (CPI) that combines seat track position, seat back angle, and measured head distance into a single scalar. When CPI exceeds a calibrated threshold, the occupant is considered to be encroaching on the far-side deployment zone, and the RCM may inhibit deployment or switch to a reduced-output profile to avoid loading the occupant's head and neck directly.

Signal Fusion and Deployment Decision

The three inputs are not evaluated independently. The RCM runs a fusion algorithm that combines crash severity (from lateral accelerometers and pressure satellites), belt status, occupant class, and position into a deployment decision. The figure below shows the signal chain from sensors through conditioning and fusion to the far-side airbag command.

The fusion logic can be expressed as a sequence of gating and threshold operations. A simplified pseudocode representation is shown below; the actual calibration constants are vehicle-specific and determined through sled and full-vehicle testing.

// Far-side deployment decision (simplified)
if (belt_status == UNBUCKLED && occupant_class == EMPTY)
    return INHIBIT;

if (occupant_class == CHILD_CRS)
    return INHIBIT;

severity = compute_severity(lateral_accel, pressure_satellite);

if (severity < threshold_low)
    return INHIBIT;

if (cpi > CPI_LIMIT)
    return REDUCED_OUTPUT;   // occupant encroaching on centerline

if (belt_status == UNBUCKLED)
    threshold_low = threshold_low * 0.85;  // earlier trigger

return NOMINAL_OUTPUT;

Practical Verification and Failure Symptoms

Verification of the sensing chain is performed on a seat rig or in a full vehicle with a programmable occupant simulator. A typical test sequence places a known load on the seat, buckles and unbuckles the belt, and moves the seat through its track range while logging the RCM's classification and CPI outputs over CAN. The expected result is that the class transitions cleanly between bands without oscillation, and that CPI responds monotonically to seat track movement.

Common failure symptoms and their likely causes include:

• Class oscillates between adjacent bands: insufficient moving-average window or a drifting tare value; recalibrate the seat empty offset.
• Belt status stuck in UNBUCKLED: open or shorted buckle switch circuit, or a Hall sensor with a missing magnet; check connector pin tension and sensor supply voltage.
• CPI does not change with seat track: track position sensor signal out of range or a CAN timeout; verify sensor output against a known seat position.
• Far-side airbag inhibits with an adult occupant: capacitive sensor detecting moisture or a seat heater fault; inspect for liquid ingress and verify the shield ground.

The key takeaway is that far-side deployment is a multi-input decision. Belt status provides the restraint context, occupant size sets the inflation energy, and position detection guards against deploying into an occupant who has already moved into the deployment zone. The RCM's fusion logic ties these together with crash severity, and the calibration must be verified across the full occupant matrix rather than a single nominal case.