• Ei tuloksia

Interface technology feasibility studies

6.2 Interface technology feasibility studies

Table 6.1 summarizes the differences of cable based (electrical) and wireless (RF) interfaces based on the work performed in this Thesis. Table 6.2 summarizes the key differences of the three main interface technologies of interest to this Thesis, Bluetooth, USB, and Zigbee. Other, more detailed results, are presented next.

USB was found to be robust and well suitable for medical applications [P1][P2], and it will likely become popular also in other measurement and data acquisition applications [P3]

[Gre09]. The complexity of USB is hidden in the host device, and peripheral devices require lighter implementation [P1]. In [P3] it is shown that USB 1.1 can be electrically isolated, although the implemented solution is not universally applicable as such. Paper [P1] deems the capacitive isolation as unusable. Although this is the case for USB data signal isolation, it still has potential for medical isolation as shown by [Pii07] (using RF modulation over capacitor) and TI (integrated high-speed digital isolator products ISO721/722). By using extremely small capacitances, of order of pF, the capacitance over the isolation barrier can be reduced to the level of traditional isolation circuitry [Pii07]. However, only very high frequencies, of the order of GHz and above, can pass through these small capacitances. Lower signal frequencies require the use of modulation technologies. Alternatively, very high frequency digital signals can be isolated using a series capacitor and latch circuit, as presented in [Kug09]. Both of these methods are unidirectional by nature, e.g., full-duplex operation requires separate circuits for both directions and circuit for detection of transmission direction. The use of these kinds of circuits is similar to optoisolators, but due to the faster operation of capacitive isolators, they are an attractive alternative in high-speed applications. In medical devices, the voltage and clearance regulations, which are easier to meet with transformers, limit their use. In case of USB, the 1.0 to 2.0 standards (low, full, and high-speed communication) are all based on 12 MHz bus clock speed, with the high-speed USB using faster microframes within the basic frames. USB 3.0 has half-duplex data signals operating at 4.8 GHz bus speed that could be isolated using capacitive isolation. However, USB 3.0 connector also includes signals for the USB 1.0 to 2.0 versions to retain backward compatibility, and therefore the same limitations apply for an overall solution as do for the older USB versions. To isolate USB data signals using capacitive isolation a fast, extremely low latency modulator with extremely fast detection of transmission direction, would be required. As has been shown by the recently released Analog Devices ADuM 4160, this kind of modulation and direction detection can be implemented with transformers currently at only low and full speed (USB 1.0-1.1). USB has been chosen as a transport method for the ISO/IEEE 11073 standard, which thanks to promotion by the Continua Alliance and the IHE, is expected to become widely adopted.

Wireless technologies can assist and improve care in hospitals [P4]. Bluetooth and other PAN technologies may not be the best technologies for hospital wide networks, but they have

Cable-based (electrical) Wireless (RF)

Power The interface can supply power. Requires battery or external power source.

Maximum range

Defined by the cable length. Can not be defined exactly, the performance will deteriorate with distance.

Security Physical access or very complicated tech-niques required to eavesdrop.

Everyone within the transmission range can pick up the signal.

Electrical safety

Electrical connection outside the device casing requires electrical isolation and pro-tection from over-currents.

No physical connector requires, the de-vice casing can be made completely non-electrical (requires batteries).

EMI caused Very small when using proper cables. Depends on transmission power.

Susceptibility to interfer-ence

Cables can pick up interference if not properly shielded and isolated. Data com-munication not easily effected.

Interference decreases communication per-formance. Small and sensitive portable sen-sory device (analog inputs) may pick up in-terference from its own transmissions.

Device mobility

Cable restricts movement. The communi-cation is not effected by movement.

No movement restrictions, the connection will deteriorate as range grows. Movement alters connection quality unpredictably and in some cases may change the underlying network topology which can effect also other devices in the premises.

Quality of service

Can be implemented in a way that band-width is guaranteed by allocating the de-vice a fixed portion of the bandwidth.

Unpredictable medium, data can be priori-tized, and time reserved for retransmissions, but bandwidth can not be guaranteed.

Bandwidth utilization

Good, only packet headers decrease effi-ciency. Increased data payload size im-proves bandwidth utilization but increases latency.

Packet headers reduce efficiency, and smaller packet sizes/payload are preferable to reduce possibility of interference and re-transmissions. Space has to be allocated for retransmission and higher bandwidth uti-lization reduces reliability.

Usability Number of attached devices increases the number of cables and degrades usability.

Number of attached devices does not di-rectly effect usability.

Topology management

Cables reflect the connections between the devices.

Visual identification and management of connections requires additional components or systems.

Global use Not restricted (electrical safety must be approved).

Use of RF bands is regulated. Only a few globally available radio bands exist.

Table 6.1: Comparison of the use of cable based and wireless technologies in medical devices from the interface and data communication viewpoint.

6.2 Interface technology feasibility studies 89

Bluetooth Zibgee USB

Type Wireless, 2.4 GHz Wireless, 2.4 GHz and 868/915 MHz

Cable based (wireless ver-sion in development)

Data rates Moderate to high Low High to very high

Signaling rate (nomi-nal)

Up to 3 Mbps using BT ra-dio, higher using alterna-tive MACs (Chapter 2)

Up to 0.25 Mbps (2.4 GHz) Up to 4800 Mbps (Chapter 2) Data link

robustness

Good, frequency hopping Moderate, narrow band Very good, shielded cable Networking Small scale piconets,

point-to-multipoint

Supports forming bigger networks and different topologies

Tiered star topology, can be extended with hubs, but all communication is be-tween host and a end-device

Power con-sumption

Moderate to low Low to very low Power can be supplied from the bus

Table 6.2: Comparison of Bluetooth, Zigbee and USB interfaces.

several uses in local monitoring applications, including both measurement signal communication and monitoring user interfaces [P4]. Bluetooth and Zigbee are the most dominant wireless PAN technologies, and operate in the radio band which allows virtually world wide coverage [P5].

Zigbee offers better networking functionality than Bluetooth and slightly lower power consump-tion due to Bluetooth’s frequency hopping. Frequency hopping used by Bluetooth makes it more robust and less susceptible to interference [P4]. New low-power Bluetooth version reduces Zigbee’s benefits in low-power applications. Bluetooth has also been chosen as a transport method for the ISO/IEEE 11073 standard. Zigbee has Health care application profile which supports ISO/IEEE 11073 device specializations, and it is likely that it too will be included as an ISO/IEEE 11073 transport method in the future.

[P5] shows that in 433 MHz to 5.9 GHz range only the 2.4 and 5.8 GHz ISM bands offer sufficient resources for a medium data rate radio link. Medium data rate in this case was defined to be in the order of 60 kbps. Ultra wide band technologies are problematic for regulators and may cause EMI when transmitting due to the large energy related to the transmission. In [P5]

we identified six issues which should be considered when selecting a wireless technology:

• Target environment - Is the use of RF bands limited, are some bands more preferable because of regulatory issues.

• The amount of integration - RF module requiring only external antenna vs. plain chip which requires board design and additional electronics.

• Physical chip/module parameters - Power consumption, operating voltage, interface to the rest of the system.

• How much does the chip/module handle the protocol - Implementing the protocol (e.g., Bluetooth higher layers) needs resources.

• Software support from the manufacturer.

• Price, availability, alternative components if the technology is discontinued.

In [P4] it was found, that PAN technologies such as Bluetooth have a clear usage area even in hospitals equipped with LAN and WLAN. The main application area was noted to be the point-of-care data communication, transfer of measurement data from the measurement instruments and other bedside devices to a bed side coordinator device, i.e., patient monitor or similar device.

Bluetooth based wireless user interfaces were also proposed. Short-range wireless interfaces can also be used as a wireless patient identifier tag, and for staff location discovery. Specific RFID technologies have later been developed for these kinds of “tagging” purposes.