Personal help button and administrator system for a low bandwidth personal emergency response system (PERS)

1. A personal help button for a personal emergency response system, PERS, comprising:

a button for activating an emergency call;

an audio device for receiving a verbal audio input (VAI) from a user and delivering a verbal audio reply (VAR) to the user;

a conversation algorithm for converting at least some of the verbal audio input (VAI) into a low bandwidth format output signal (LBFOS) and for converting a received low bandwidth format reply (LBFR) into a verbal audio reply (VAR), wherein the conversation algorithm comprises a trained neural network and the low bandwidth format has a bandwidth below 500 bits/second; and

a transmission and reception system for the low bandwidth format, for transmitting a low bandwidth format output signal to a remote PERS helper and receiving the low bandwidth format reply from the remote PERS helper.

2. A personal help button as claimed in claim 1 , further comprising:

a speech synthesizer for generating the verbal audio reply.

3. A personal help button as claimed in claim 1 , wherein the low bandwidth system has bandwidth below 250 bits/second.

4. A personal help button as claimed in claim 1 , wherein the low bandwidth system comprises narrowband IoT, Lora, Narrowband IoT, Wi-Fi HaLow, LTE Cat M1 or Sigfox technology systems.

5. A personal help button as claimed in claim 1 , wherein the trained neural network is trained using verbal audio data from previous emergency events.

6. An administrator system for a personal emergency response system, PERS, comprising:

an input for receiving an emergency call from a user;

a transmission and reception system for receiving a low bandwidth format input signal (LBFIS) from the PERS user and for transmitting a low bandwidth format reply signal (LBFRS) to a PERS user;

a microphone for receiving a verbal audio reply (VAR) from a PERS helper; and

a conversation algorithm for converting the low bandwidth format input signal (LBFIS) into a verbal or textual communication for the helper, and for converting a verbal audio reply (VAR) from the helper to the low bandwidth format reply signal (LBFRS), wherein the conversation algorithm comprises a trained neural network and the low bandwidth format has a bandwidth below 500 bits/second.

7. An administrator system as claimed in claim 6 , further comprising a speech synthesizer for converting the verbal or textual communication into a verbal audio signal (VAS).

8. An administrator system as claimed in claim 6 , wherein the low bandwidth system has bandwidth below 250 bits/second.

9. An administrator system as claimed in claim 6 , wherein the low bandwidth system comprises narrowband IoT, Lora, Narrowband IoT, Wi-Fi HaLow, LTE Cat M1 or Sigfox technology systems.

10. An administrator system as claimed in claim 6 , wherein the trained neural network is trained using verbal audio data from previous emergency events.

11. A non-transitory computer-readable medium implementing a computer program comprising computer program code means which is adapted, when said program is run on a computer, to implement a communication method for a user of a personal help button of a personal emergency response system, PERS, to communicate with a helper of the PERS administrator, the method comprising:

activating an emergency call;

receiving verbal audio input from the user;

using a trained neural network conversation algorithm to convert at least some of the verbal audio input into a low bandwidth format output signal;

transmitting the low bandwidth output signal to the helper over a channel with a bandwidth below 500 bits/second;

receiving a low bandwidth format reply from the helper; and

using the conversation algorithm to convert the low bandwidth format reply into a verbal audio reply.

12. A communication method for a helper of personal emergency response system, PERS, administrator to communicate with a user of the PERS:

receiving an emergency call;

receiving a low bandwidth format input signal from the user over a channel with a bandwidth below 500 bits/second;

using a trained neural network conversation algorithm to convert the low bandwidth format input signal into a verbal audio signal;

using the conversation algorithm to convert a verbal audio reply into a low bandwidth format reply signal; and

transmitting the low bandwidth format reply signal to the user.