Reaction strategies

This reaction strategy document is motivated by the fact that different error codes have different implications and severity and therefore have to be treated differently.

Severity levels definitions

Severity	Meaning
Information	This level is used to convey information to the user. No persistent error has occurred. No action needs to be taken by the user. The system remains in full operational state.
Warning	This level indicates that an error was detected but is not serious enough to interfere with the running of the system and potential embedded applications. Performance of the system may be reduced. Please see the handling strategies section for more details on how fix the specific issue.
Severe Error	This level indicates that a serious error was detected and is persistent. This inhibits the full operability of the system and needs to be fixed by the user to ensure the full feature level and system performance. In comparison to Unrecoverable error cases, the system doesn’t shut off the respective components automatically . The error may be fixed in live operation of the system by the user. Please see the handling strategies section for more details on how fix the specific issue.
Unrecoverable	This level usually indicates an error that forces termination of respective embedded process. In consequence in most cases at least a reboot is required to reset the respective components. Additional preliminary solutions steps preceding the reboot may be required. Please see the handling strategies section for more details on how fix the specific issue.

Severity level dependent actions

Please verify by the error code severity level if this error inhibits the operation of the device. Act accordingly as described in the handling strategy.

Boot-Sequence

ID	Name	Description	Severity level	Handling strategy
101014	ERROR_BOOT_SEQUENCE_VPU_EEPROM	VPU EEPROM content invalid	Severe	VPU specific EEPROM content can not be read. reboot if persistent factory reset
101018	ERROR_BOOT_SEQUENCE_INVALID_CONFIGURATION	Unable to construct a valid device configuration	Unrecoverable	Bring-up failed and unrecoverable. Manual reconfiguration required. reboot if persistent factory reset

ID	Name	Description	Severity level	Handling strategy
101006	ERROR_BOOT_SEQUENCE_PORT_CALIBRATION	The port calibration is invalid	Severe / Unrecoverable	reboot if persistent factory reset
101007	ERROR_BOOT_SEQUENCE_PORT_INVALID_CONFIGURATION	The init configuration does not match the port	Severe	The configuration saved on the device (saveInit) does not match the connected hard. A factory reset to erase the persistent configuration is required.
101015	ERROR_BOOT_SEQUENCE_PORT_DUMMY_CALIBRATION	A dummy calibration is used for the port	Severe	A dummy calibration is used during runtime. Measurement performance and accuracy is reduced. reboot if persistent after factory reset the hardware needs to be replaced.
101017	ERROR_BOOT_SEQUENCE_PORT_CONFIGURATION_TIMEOUT	The imager did not respond to the frame software	Severe	Camera head did not respond within 5 sec after boot-up. reboot if persistent over runtime and re-flash FW. A hardware replacement of the head might be required.
101019	ERROR_BOOT_SEQUENCE_PORT_EEPROM_OVERRIDE	A user-provided EEPROM file is used instead of the actual EEPROM	Severe	User specified EEPROM content used. Measurement performance and accuracy is reduced. reboot. If persistent after factory reset the hardware needs to be replaced.
101020	ERROR_BOOT_SEQUENCE_PORT_CALIBRATION_OVERRIDE	A user-provided calibration file is used	Severe	User specified calibration used. Measurement performance and accuracy may be reduced. reboot. If persistent after factory reset the hardware needs to be replaced.
101021	ERROR_BOOT_SEQUENCE_PORT_IDENTIFICATION	Port identification data is invalid	Severe	Camera head can not be identified. reboot.

ID	Name	Description	Severity level	Handling strategy
101000	ERROR_BOOT_SEQUENCE_TCU_INVALID_FW	The installed TCU firmware is incompatible	Severe / Unrecoverable	TCU firmware is not compatible / was not updated automatically to the correct FW dependent version reboot if persistent re-flash FW.

ID	Name	Description	Severity level	Handling strategy
101004	ERROR_BOOT_SEQUENCE_HEAD_INVALID_DRIVER	The appropriate driver for the connected head is not available	Unrecoverable	The camera head hardware may not be supported by this VPU and firmware. Disconnect all camera heads reboot and connect one camera head at a time to debug the problem.
101008	ERROR_BOOT_SEQUENCE_HEAD_INVALID_SERIALNUMBER	The serial number of an extrinsically calibrated head does not match; please recalibrate after changing heads	Unrecoverable	The extrinsic calibration saved on the device (saveInit) does not match the connected hard at this specific port. Factory reset to remove all the persistently saved settings or debug the configuration manually for each port.
101016	ERROR_BOOT_SEQUENCE_HEAD_INVALID_COMBINATION	Only 2D/2D pairing or 3D/3D pairing is allowed	Unrecoverable	Different camera streams connected to the same deserializer. Update hardware connectivity and reboot

ID	Name	Description	Severity level	Handling strategy
101010	ERROR_BOOT_SEQUENCE_HEAD_ICC_FW_CHECK	Due to incorrect ICC firmware data an ICC update is attempted	Warning	The illumination controller FW does not match the one required by the device FW. Wait for the ICC FW update to complete. If persistent over > 5 min reboot. If persistent over reboots re-flash FW.
101011	ERROR_BOOT_SEQUENCE_HEAD_ICC_FW_UPDATE	The ICC firmware data is still incorrect after an ICC update
101012	ERROR_BOOT_SEQUENCE_HEAD_ICC_FW_INCOMPATIBLE	Due to an incompatible ICC firmware version an ICC update is attempted	Warning	The illumination controller FW does not match the one required by the device FW. Wait for the ICC FW update to complete. If persistent over > 5 min reboot. If persistent over reboots re-flash FW.
101011	ERROR_BOOT_SEQUENCE_HEAD_ICC_FW_UPDATE	The ICC firmware data is still incorrect after an ICC update
101013	ERROR_BOOT_SEQUENCE_HEAD_ICC_FLASH	Unable to read the ICC flash	Unrecoverable	The illumination controller FW does not match the one required by the device FW. reboot. If persistent over reboots re-flash FW.
101011	ERROR_BOOT_SEQUENCE_HEAD_ICC_FW_UPDATE	The ICC firmware data is still incorrect after an ICC update	Unrecoverable	The illumination controller FW does not match the one required by the device FW. reboot. If persistent over reboots re-flash FW.

ID	Name	Description	Severity level	Handling strategy
101002	ERROR_BOOT_SEQUENCE_IMU_INVALID_HW	The installed IMU (if any) was not recognised	Unrecoverable	Detected IMU hardware is not compatible. reboot.
101003	ERROR_BOOT_SEQUENCE_IMU_INVALID_CALIBRATION	The IMU calibration is either missing or incorrect	Unrecoverable	Invalid IMU calibration used at boot-up - re-flash FW

VPU

ID	Name	Description	Severity level	Handling strategy
103002	ERROR_VPU_OVERVOLTAGE_24V	Overvoltage was detected for the 24V line	Warning	Fix power supply
103003	ERROR_VPU_UNDERVOLTAGE_24V	Undervoltage was detected for the 24V line	Unrecoverable	This error may cause undefined system states. Fix power supply and reboot
103004	ERROR_VPU_OVERVOLTAGE_5V	Overvoltage was detected for the 5V line	Warning	Fix power supply
103005	ERROR_VPU_UNDERVOLTAGE_5V	Undervoltage was detected for the 5V line	Unrecoverable	This error may cause undefined system states. Fix power supply and reboot
103006	ERROR_VPU_OVERVOLTAGE_1_8V	Overvoltage was detected for the 1.8V line	Warning	Fix power supply
103007	ERROR_VPU_UNDERVOLTAGE_1_8V	Undervoltage was detected for the 1.8V line	Unrecoverable	This error may cause undefined system states. Fix power supply and reboot
103008	ERROR_VPU_OVERTEMPERATURE	Detected overtemperature for VPU main board	Severe	Provide adequate (passive) cooling via heat conduction and convection. The systems CPU and GPU performance will be throttled until adequate temperatures are reached.
103009	ERROR_VPU_NTP_NOT_SYNCHRONIZED	Unable to synchronise time with NTP server	Warning	Check NTP server availability and correct configuration. A reboot may be required to avoid time jumps in live operation due to syncing of clocks. The NTP time will be directly applied at next start-up.
103010	ERROR_VPU_TRIGGER_OVERRUN	A trigger overrun was detected	Severe	The trigger signal are not consistent: data may be lost due to lost trigger signals. Verify software trigger sending units on your IPC.
103011	ERROR_VPU_NO_FREE_IMAGEBUFFERS	More image buffers are needed than are available	Warning (sporadic) /Severe (permanent)	The data is not send fast enough. In consequence the internal queues have filled up. This may be because of internal application streams switching or because of a slow receiver. If persistent please verify network transmission speeds: GBit/s is required. Verify the receiver is adequate to receive the data frames fast enough via TCP/IP.
103012	ERROR_VPU_WATCHDOG_TIMEOUT	A watchdog timeout occurred; the component was restarted	Severe	Internal data evaluation processes have not responded in time and are restarted. The processes have been restarted but data frames may have been lost in the meanwhile. This diagnostic element is persistent (dormant) on the log after detection. To reset the logs completely a reboot is required. If the user experiences this problem more than once, please send a trace of the systems logs via email to support.efector.object-ident@ifm.com. The system trace can be retrieved via the ifm3d API and ifm3d API’s CLI.

ID	Name	Description	Severity level	Handling strategy
103000	ERROR_VPU_COMM_TCU	Unable to communicate to the TCU	Unrecoverable	The internal trigger mechanism is disturbed and image acquisition is not triggered reliably. Reboot
103001	ERROR_VPU_COMM_TEMPSENSOR	Unable to communicate to the temperature sensor	Warning	The internal communication to the VPUs temperature sensor is disturbed. Temperature values might not be accurate. If persistent reboot

Distance image processing

ID	Name	Description	Severity level	Handling strategy
102015	ERROR_DI_UNEXPECTED_CALIB_CONTENT	The port calibration contains unexpected values	Unrecoverable	Sanity checks of the 3D calibration file failed. The data acquisition of the respective camera is blocked. This is most likely due to old / incompatible hardware. If error is persistent factory reset and re-flash FW or replace hardware.
102016	ERROR_DI_MOTION_COMP_NO_EGO_DATA	Expected motion compensation data, but none was received	Warning	If motion compensation is activated but can not be performed due to missing ego data. The motion compensation feature is available only in combination with a live ODS application. Reset the port parameters to default to disable motion compensation.
102017	ERROR_DI_MOTION_COMP_EGA_DATA_TIMESTAMP_MISMATCH	Received motion compensation data is too old or contains implausible timestamps	Warning	If motion compensation is activated but can not be performed due to ego data unavailability or outdated timestamps. The motion compensation feature is available only in combination with a live ODS application. Reset the port parameters to default to disable motion compensation.

Camera head

ID	Name	Description	Severity level	Handling strategy
102000	ERROR_HEAD_OVERTEMPERATURE_ICC	Head ICC has overtemperature	Severe	Provide adequate (passive) cooling via heat conduction and convection. The camera head will not be useable until it has cooled down to operating temperature levels again.
102001	ERROR_HEAD_BROWNOUT	Detected head brownout	Severe	Fix power supply - if active / persistent reboot to reset / reboot heads.
102002	ERROR_HEAD_ENVELOPE_PAUSE_VIOLATION	Detected violation of the envelope pause ratio	Severe	If detected the system will enter an error state and block the image acquisition - to reset, reboot.
102003	ERROR_HEAD_OVERVOLTAGE	Head has overvoltage	Severe	The respective head will be blocked from acquiring images until no overvoltage is present - fix the power supply and cables.
102004	ERROR_HEAD_UNDERVOLTAGE	Head has undervoltage	Severe	The respective head will be blocked from acquiring images until no undervoltage is present - fix power supply and cable hardware. A reboot may be required to exit the error state.
102005	ERROR_HEAD_VCSEL_OVERCURRENT	Head VCSEL has overcurrent	Unrecoverable	The active illumination has shut off. The respective head will be blocked from acquiring images until the system is rebooted
102006	ERROR_HEAD_EYESAFETY_SHUTDOWN	Head was shut down due to eye-safety concerns	Unrecoverable	The active illumination has shut off due to a eye safety violation. The respective head (including its illumination) will be blocked until system is rebooted
102007	ERROR_HEAD_INVALID_CALIBRATION	Head current calibration is invalid	Unrecoverable	The respective head has detected a check sum error for its calibration - will be blocked until system is rebooted if persistent factory reset, finally re-flash FW and/or replace hardware.
102008	ERROR_HEAD_ENVELOPE_PULSE_ERROR	Detected pulse error for envelope signal	Unrecoverable	The active illumination has shut off due to a eye safety violation. The respective head will be blocked until system is rebooted
102009	ERROR_HEAD_VCSEL_OVERTEMPERATURE	Head VCSEL has overtemperature	Unrecoverable	The active illumination has shut off. The respective head will be blocked until system is rebooted
102010	ERROR_HEAD_ENVELOPE_MAX_DURATION	Maximum length of envelope signal was detected	Unrecoverable	The active illumination has shut off due to a eye safety violation. The active illumination has shut off. The respective head will be blocked until system is rebooted
102012	ERROR_HEAD_IMAGER_RESET	The 3D imager was unexpectedly reset	Severe	The respective imager process was reset. Data frames may be lost in the meanwhile.
102019	ERROR_HEAD_VCSEL_OVERVOLTAGE	Head VCSEL has overvoltage	Unrecoverable	The respective head will be blocked from acquiring images until no overvoltage is present - fix the power supply and cables.
102020	ERROR_HEAD_ENVELOPE_PAUSE_CURRENT	Detected a violation of the envelope pause current	Unrecoverable	The active illumination has shut off. The respective head will be blocked from acquiring images until the system is rebooted
102021	ERROR_HEAD_VCSEL_SHUTDOWN	Head VCSEL was shut down; VPU must be rebooted to recover	Unrecoverable	The active illumination has shut off. The respective head will be blocked from acquiring images until the system is rebooted

IMU

ID	Name	Description	Severity level	Handling strategy
104000	ERROR_IMU_DATA_IMPLAUSIBLE	Received implausible IMU data	Severe	The system checks for IMU timestamp and temperature value ar implausible. This warning is active until new plausible data is available. If error is persistent over longer time durations reboot. If error persists over reboots in long-term system operation the hardware may be corrupted. A VPU hardware replacement may be required.

Hardware port

ID	Name	Description	Severity level	Handling strategy
102011	ERROR_PORT_FRAME_TIMEOUT	The 3D imager detected a frame timeout	Warning (sporadic) / Severe (permanent)	Data may have been lost, the imager process has recovered. If persistently active reboot
102013	ERROR_PORT_FPDLINK	An unrecoverable FPD-Link error occurred	Severe / Unrecoverable	reboot if persistent check for hardware defects
102014	ERROR_PORT_ALGO_INTERNAL	Internal error in the port algorithm	Warning (sporadic) / Severe (permanent)	If persistent reboot for non-series hardware check internal Algorithmic errors. If the user experiences this problem more than once, please send a trace of the systems logs via email to support.efector.object-ident@ifm.com. The system trace can be retrieved via the ifm3d API and ifm3d API’s CLI.
102018	ERROR_PORT_MODE_ALGO_INCOMPATIBLE	The appropriate algorithm for the selected mode is not available	Unrecoverable	If persistent reboot. If persistent re-flash FW

ODS application

ID	Name	Description	Severity level	Handling strategy
105000	ERROR_ODSAPP_EXTR_VPU_CALIB_IMPLAUSIBLE	Implausible extrinsic VPU calibration	Severe	The app checks plausible IMU / VPU extrinsic calibration. Perform a IMU / VPU calibration and sanity check link. For more details see the chapter Implausible extrinsic calibrations
105001	ERROR_ODSAPP_IMAGE_IMU_DELAY_IMPLAUSIBLE	Detected an implausible delay between IMU and image data	Warning (sporadic) / Severe (persistent)	The app checks plausible 3D data timestamps. No internal escalation is performed if active. If persistent reconfigure the ODS application / reboot. If the user experiences this problem more than once, please send a trace of the systems logs via email to support.efector.object-ident@ifm.com. The system trace can be retrieved via the ifm3d API and ifm3d API’s CLI.
105002	ERROR_ODSAPP_VO_EXTR_DI_CALIB_IMPLAUSIBLE	Implausible extrinsic calibration of visual odometry head	Warning	The app checks camera specific calibration for plausibility. If this error is active the respective camera can not be used in this app for estimating the systems odometry. If the camera is setup to see the floor and the diagnostic is persistent verify the mounting setup and extrinsic calibration values. See the additional comments in camera head calibrations and implausible extrinsic calibrations.
105003	ERROR_ODSAPP_PARAMETER_PLAUSIBILITY_CHECK_FAILED	Implausible ODS configuration parameters	Severe	The app checks its configuration for plausibility. If implausible the app is set to error state and has to be reconfigured to exit error state. Check the app specific JSON schema for configuration verification.
105004	ERROR_ODSAPP_VO_IMAGE_FRAMERATE	Insufficient framerate of ODS input images	Warning (sporadic) / Severe (permanent)	The app checks plausible 3D frame timestamps. No internal escalation is performed if active. Sporadic diagnostics are expected. If this error is permanently active then system performance may be severely reduced. To fix the issue, reboot / re-flash. Depending on the additional load from OEM Docker containers, these errors may be activated. If the error persists, please reduce your system load (especially CPU and GPU) to ensure uninterrupted ODS output.
105005	ERROR_ODSAPP_INTR_IMU_CALIB_IMPLAUSIBLE	Implausible intrinsic IMU calibration	Warning	The app checks performs repeated IMU standstill calibration during live operation. If the estimated calibration values differ more than the internal threshold from their expected values this error will be activated. This is an internal error - no internal escalation strategy will be applied. It will be automatically be reset by the application at the next possible instance. Please check the suggestions Mounting setup. If the error persists and has only occurred during operation, the IMU may be defective: please verify IMU functionality using the provided IMU extrinsic calibration verification strategy.
105006	ERROR_ODSAPP_UNSTABLE_FRAMERATES	Unstable framerate of at least one input stream	Warning (sporadic) / Severe (permanent)	At least one sub-component of the app is unstable due to unstable framerate of the input data stream. See ODS unstable framerate
105007	ERROR_ODSAPP_VELOCITY_UNAVAILABLE	Unable to determine velocity	Warning	The app performs ego-motion estimations. No internal escalation is performed if active. Mounting setup.
105008	ERROR_ODSAPP_DEFAULT_ZONE_USED	Using the default ODS zones	Info	The app checks for default zone configuration. No internal escalation is performed if active. To silence this warning change the setting of /applications/instances/appX/configuration/zones/zoneConfigID
105009	ERROR_ODSAPP_STANDSTILL_CHECKS_NOT_EXECUTED	Standstill condition not detected	Warning / Severe	The app failed to perform ego-motion estimations. If the camera is not setup to see the floor, this information can be disregarded. An initial standstill period of about five seconds after the O3R ODS application is mandatory, since plausibility checks must be executed and IMU biases have to be estimated before ego motion can become available. This error is set on initialization of the O3R ODS application. It is reset after plausibility checks have been executed once. If the error persists even though the vehicle is at standstill and a large portion of the floor is visible for the camera assigned for visual odometry, this is an indication that there is an issue either with the IMU or with the extrinsic camera calibration.
105011	ERROR_ODSAPP_EXTR_DI_CALIB_IMPLAUSIBLE	Implausible extrinsic head calibration	Severe	The app checks plausible extrinsic head calibration. The O3R ODS app is functional in this state. The extrinsic calibration of the ODS ports must be fixed before new application data will be send. A RUN-CONF-RUN state change of the application may be required to reinitialize the application. Check your mounting setup.
105012	ERROR_ODSAPP_TEST_MODE_WARNING	A protected parameter was set to enter a test mode	Severe	The app is configured with development features active, for example the overwriting of velocity information status. Such behavior is not expected on devices in operation. To fix this issue please reboot, re-flash. If the error persists after a re-flash please reach out to ifm at support.efector.object-ident@ifm.com.
105013	ERROR_ODSAPP_ERROR_ODSAPP_FOV_INSUFFICIENT_FOR_NEGATIVE_OBSTACLES	FOV is insufficient for negative obstacle detection	Severe	The app detects insufficient amount of floor. No negative obstacles can be detected - the ODS performance is reduced. Please verify an active negative obstacle detection configuration under /applications/instances/appX/configuration/portX/negObst/enableNegativeObstacles. A different mounting setup may be required to enable the negative obstacle detection feature.

Handling strategies description

Reboot / power cycle

A software reboot can be performed via the ifm3d library through the CLI, Python or c++, or using /sbin/reboot on the VPU.

Alternatively one can power cycle the device: the power connection should be interrupted for at least 5 seconds.

Factory reset

The device can be reset to factory conditions and configuration via the ifm3d CLI, Python or c++.

Flash firmware anew

Update the firmware again. For backward compatibility, see the firmware release notes.

Camera head connectivity

Different camera head imagers, for example 2D RGB or 3D TOF imagers of different resolutions, need to be connected to separate deserializers. Ports [0,1], Ports [2,3], Ports [4,5] are paired to the same deserializer internally and can therefore only handle image streams of the same type. Checks can be performed by verifying the cable description at the mini-FAKRA connector or via the complete JSON system configuration.

Image loss due to trigger overruns

The system runs in a triggered mode internally for “RUN” and “IDLE” state. The difference is: for RUN state the system gets triggered continuously, whereas in “IDLE” state, only single software trigger signals are propagated.

The trigger overrun only applies to the software triggered state “IDLE,” in case multiple trigger signals have been sent and cannot be accepted by the system. In case of trigger overruns the system may have lost trigger signals, which means the respective images may not have been acquired or their frame start may have been delayed by an undefined time. Additionally check for network latency which cause trigger signals to get delayed.

Image buffer availability

Internally, the system monitors its data buffer states. Each data stream is buffered into its own buffers with an anticipated buffer size, depending on the type and data frequency.

When the data is not retrieved fast enough or the queue is not popped often enough, an error is activated:

1. Externally pointing data streams: the data retrieval rate is not adequate. Possible bottlenecks are: network bottlenecks (insufficient network speed / latency ), OEM application running at low data retrieval rates, etc. Check your software and hardware for bottlenecks.

2. Internally pointing data stream: an ifm applications on the device is not processing data with the required frequency. See the application specific error codes. Check for additional (Docker based) applications which have high resource requirements.

For case 1 (external): in case data is not retrieved fast enough the data buffer will be partially flushed. For case 2 (internal): in case data is not retrieved fast enough the data acquisition framerate can be slowed down.

Watchdog errors or timeouts

The system monitors its internal process states via watchdogs. Each watchdog has it own respective timeout, which is configured not to activate under normal operation conditions.

High processing requirements or certain re-configurations during runtime can cause sporadic triggers of watchdog timeouts. This can be tracked by dormant and low error counts. This case can be neglected.

In case repeated active watchdog errors and high error counts increases (> 5 per min) the system may not be self recoverable and requires a reboot.

FPD-Link errors

The system monitors the FPD-Link communication between the VPU and each camera head.

The FPD-Link communication can perform self healing actions for most cases. If ERROR_PORT_FPDLINK errors are active the system entered a non-recoverable FPD-Link dependent error state. Possible sources of such errors are loose cable connections, continuous EMV / ESD disturbances on the cables, damaged cables, etc. Check for possible hardware malfunctions first and replace FAKRA cables to ensure no hardware defects are present.

Internal error in the port algorithm

The algorithmic data evaluation pipeline on the embedded device monitor for uncaught internal error cases. These should not be present for series hardware and software as supplied by ifm.

For pre-release software and prototype hardware this error may occur. Please check the release notes for known defects for your specific non-series software and hardware. For older hardware that is no longer supported by newer firmware or software releases, this error may occur. In these cases the data will most likely still be available but may be reduced in performance and accuracy.

Mounting setup

Depending on the application, specific mounting setups may be required to ensure all required data appears in the FOV of the respective camera head.

Motion compensation

The motion compensation algorithm requires the floor to be part of the scene to work. If the floor is not part of the scene or can not be identified as floor for longer durations (> 50 frames), motion compensation will be deactivated. Fix:

• Double check your mounting setup for rigidity.

• Check for hardware cleanliness: dust and fingerprints on the housing will reduce performance and possible floor range.

• If hardware has recently been replaced check for correct alignment and extrinsically calibrate the system.

Camera head extrinsic calibrations

The camera head specific extrinsic calibration errors are set if the extrinsic camera calibration of the camera assigned for visual odometry is implausible. To this end, the intersection between the camera’s field of view and the floor plane is considered. The minimum radial distance of this intersection must not be larger than 1.0 meters, while the maximum radial distance must be larger than 1.5 meters. This check is only executed once at startup of O3R ODS application.

Implausible extrinsic calibrations

Certain implausible extrinsic calibrations (Camera heads or VPU to RCF) can be detected by the embedded applications. The respective diagnostic messages are activated.

The plausibility checks regarding the extrinsic camera calibration and the extrinsic VPU calibration are based on the following assumptions (identical to those of the O3R ODS application) regarding the definition of the user coordinate system: In the user coordinate system, the ground plane must be z=0. The z-axis of the user coordinate system must point up. The user coordinate system must be right-handed. It is also assumed that gravity acts along the z-axis of the user coordinate system (at least during the initial standstill after starting the O3R ODS application).

VPU: This error is set when the extrinsic VPU calibration is implausible. The extrinsic VPU calibration is considered plausible if the angle between the measured acceleration and the expected acceleration, that is gravity acting along the z-axis of the user coordinate system, does not exceed five degrees. This check is performed only at rest. Once the extrinsic VPU calibration has been verified as plausible, this check is no longer performed until the O3R ODS application is restarted.

Check the setting in /applications/instances/appX/configuration/vo/voPorts and remove the incorrect port from the list. Please note that the performance of O3R ODS applications will be degraded if ego-motion cannot be estimated. Such configurations are only supported if the AMR is running at slow speeds (< 0.5 m/s).

Camera heads: Since visual odometry is obtained from image measurements in the ground plane, the plausibility check of the extrinsic camera calibration (or rather the suitability for visual odometry) considers the intersection between the camera’s field of view and the ground plane. The minimum radial distance of this intersection must not be greater than 1.0 meters, while the maximum radial distance must be greater than 1.5 meters. If this is not the case, ERROR_ODSAPP_VO_EXTR_DI_CALIB_IMPLAUSIBLE is set in O3R diagnostics. This check is performed only once at the start of the O3R ODS application.

Floor structure: In addition, visual odometry requires the floor to have a minimum amount of structure or texture. In addition to the actual texture of the floor (such as aggregate in concrete), lane markings, abrasions, or scratches usually provide texture to the floor.

Ego-motion estimation based on IMU measurements and visual measurements

The internal visual odometry is based on two data sources: IMU measurements and visual measurements. It may be unavailable in cases of insufficient texture on the floor, ambient light, for example sun light, or if only small portions of the floor are visible, for example when an object covers a large part of the field of view. Short-term outages of visual odometry signal, while the vehicle is still moving (up to tens of seconds) can be overcome in most cases. In this case, ego motion becomes available again shortly after the first valid visual odometry updates. Long-term outages (of several minutes) require the vehicle to be at standstill for > 5 sec for the ego motion to become available again and the respective active diagnostics to be reset.

ODS unstable framerate implications

In case an sub-component unstable framerate has been detected the system runs with reduced performance. This includes:

• throttled input data stream framerate

• reduced output performance or data quality.

This error can be caused by high resource load (CPU / GPU) scenarios on the embedded device. If additional resource heavy software components are run simultaneously this can cause system overload.

Solution strategy:

• Pin user processes started from an OEM Docker container to the ARM A57 cores of the TX2 unit (by using a taskset of 0x39). Note that the DENVER cores of the TX2 unit are reserved for the ODS application and can not be used by OEM user application.

• Reduce the number of simultaneously active cameras (see /applications/instances/appX/configuration/activePorts)

• Reduce workload of OEM Docker containers by optimizing for runtime or offloading to other processing hardware.

Ethernet connection issues

Ethernet connection issues are as of FW 1.0.14 not tracked as diagnostic events or errors. A Ethernet patch cable disconnect for example will not result in an error state as indicated by the status LEDs or diagnostic entries. Such an error can only be tracked using the ifm3d trace call to retrieve the full embedded systemD journal log.